A startup changes the model behind its AI feature. The new model is faster, cheaper and performs better on public benchmarks. The engineering team runs its tests, deploys the update and waits for the improvement.

Instead, support tickets begin to arrive. The assistant is less accurate on short questions. It misunderstands customers who mix languages. Nothing is completely broken, but the product is noticeably worse.

This is one of the difficult realities of building with generative AI: a system can pass every conventional software test while its behaviour deteriorates. The cheapest protection is not another monitoring platform or a more powerful model. It is a small, carefully maintained collection of real examples known as a golden dataset.

AI regressions are different

Traditional software is usually tested against predictable outputs. Give a function a particular input and it should return an exact result.

Generative AI does not work that way. Two answers can use entirely different words and still be equally correct. Conversely, an answer can sound fluent and professional while omitting a critical fact or inventing something the system does not know.

This makes informal testing dangerously attractive. Someone tries five prompts in a staging environment, reads the answers and concludes that the new version “looks better”.

That judgement becomes unreliable as the product grows. A support assistant serving several Southeast Asian markets may encounter English, Thai, Vietnamese and code-switching within the same conversation. Users submit fragments, screenshots, misspellings and requests that the product team never anticipated.

The polished questions used in demonstrations rarely resemble this traffic. A golden dataset makes quality visible. It contains representative user inputs together with a description of what a successful response must and must not do. The team runs the same examples whenever it changes a model, prompt, retrieval pipeline or tool.

It does not replace production monitoring, user feedback or A/B testing. It catches problems before those slower signals arrive.

Begin with real behaviour

The first version does not need thousands of examples. Thirty to fifty inputs from actual usage are enough to begin testing one important product behaviour.

For a customer-support assistant, that behaviour might be answering refund questions, for a financial product, it might be explaining a transaction without offering unauthorised advice, and so on.

Production examples are far more valuable than questions invented during a workshop. They contain the ambiguity, incomplete information and unusual phrasing that expose weaknesses in the system.

Also Read: SEA’s AI boom has a water problem it cannot offset away

Naturally, using production traffic requires appropriate consent, access controls, retention rules and removal of personal information. If the product has not launched, examples from internal testing or a closed beta can be used temporarily. Synthetic examples are useful for getting started, but they should gradually be replaced by real interactions.

The objective is not to construct a perfect benchmark. It is to capture a small but recognisable sample of how people actually use the product.

Test failures, not just the happy path

Many teams create evaluation sets that resemble product demonstrations: clear questions, correct terminology and complete information. Unsurprisingly, the system performs well.

A useful dataset should contain the situations most likely to cause damage. These might include ambiguous requests, unsupported languages, missing account information, contradictory documents, attempts to override instructions or questions that should be escalated to a person. Support tickets and user complaints are often the best source of such examples.

Every production failure should leave something useful behind. Once the immediate problem has been resolved, the interaction should become a new evaluation case. That ensures the same class of failure is less likely to return quietly after the next update.

Over time, the dataset becomes a record of what the team has learned about its users and its product.

Evaluate outcomes, not identical wording

For open-ended AI outputs, requiring an exact answer is usually the wrong approach.

Consider a user asking for a refund. A successful response might need to mention the refund policy, correctly identify the order, avoid promising an outcome and call the account lookup tool before answering. Many different responses could satisfy those requirements.

The evaluation should therefore describe the outcome:

• What information must appear?
• What must never appear?
• Which tool or source must the system use?
• When must the request be escalated?
• Are there limits on length, latency or cost?

Also Read: Thailand’s mobility future will be decided by data, not just vehicles

Some checks are inexpensive and objective. A test can verify whether the response contains a required fact, follows a defined structure, calls the correct tool or stays within a token limit.

Subjective qualities such as tone or clarity may require human review or another model acting as a judge. But expensive AI-based scoring should not be the default. Use the cheapest test capable of detecting the problem.

Both OpenAI’s evaluation guidance and Anthropic’s work on agent evaluations emphasise structured, task-specific evaluation rather than relying on general benchmarks alone.

Keep the dataset small enough to trust

A dataset becomes useless when it is too large for anyone to inspect. Each regression report should show the affected example, the previous response, the new response and the reason it failed. A dashboard announcing that quality fell from 84 to 81 per cent is not enough. Engineers need to see what became worse.

The dataset should also be versioned alongside the product. When expected behaviour changes, the reason should be recorded. A small portion of the examples can be held back from everyday development so the team does not unconsciously tune the system only to the cases it sees.

Finally, someone must own the dataset. Shared responsibility often means no responsibility. The owner should add newly discovered failure modes, remove obsolete examples and ensure evaluations continue to run as the product changes.

A practical starting point

A startup can establish a useful regression process within a week. Choose the single AI behaviour that matters most to customers. Collect 30 to 50 representative inputs. For each one, write down two to four conditions that define a successful outcome. Automate the inexpensive checks and run them whenever the relevant system changes.

AI teams will never eliminate uncertainty completely. Models change, products evolve and users find new ways to surprise us. But a team should always be able to answer one basic question before releasing an update: did this make the product better or worse? A golden dataset is the cheapest reliable way to find out.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post The cheapest way to stop your AI product from regressing appeared first on e27.

In Southeast Asia’s heat and humidity, keeping fruit, vegetables and ready-to-eat food fresh is not just a logistics problem but an economic one, too. Produce can lose value at every stage of the chain, from farm to wholesaler to supermarket shelf, while restaurants and households often throw away food long before it should have spoiled.

Singapore-based N&E Innovations is betting that part of the answer can come from the waste stream itself.

Also Read: Deeptech’s secret: Ignore the market, master the engineering, and let opportunity find you

The deeptech company has raised approximately US$1.7 million in Series A funding led by Australian agrifood investor Tundra Capital. SGInnovate, The Radical Fund, Archipelago VC and SG7 Group also joined the round, alongside existing backers Cercano, SEEDS Capital, Elev8 Capital and Qian Hu Corporation.

Founded in 2020 by biomedical scientist Didi Gan, N&E has developed ViKANG99, a patented edible antimicrobial ingredient made from upcycled agricultural by-products, including discarded cashew nut husks. The company says the ingredient can help fresh produce last up to four times longer by slowing the growth of bacteria and mould.

“Food waste is usually seen as something we need to get rid of. We see it as a resource,” Gan said. “We can take something like a discarded cashew nut husk, extract the compounds that naturally fight microbes and turn them into an ingredient that can help protect food.”

From cashew husks to cling wrap

ViKANG99 works by extracting naturally occurring antimicrobial compounds from agricultural waste, refining them into a food-grade ingredient, and embedding or applying that ingredient across different use cases.

One of N&E’s first commercial products is The Orange Wrap, which the company describes as the world’s first antibacterial cling wrap. Conventional cling film acts largely as a barrier. N&E’s version incorporates ViKANG99 into the material, allowing it to actively inhibit microbial growth on food surfaces.

The same core ingredient is also being used in KeepWell, N&E’s plant-based cleaning and hygiene range, and in post-harvest agriculture products that can be applied to fruit and vegetables after harvest. These include a post-harvest wash and guard system designed to suppress mould and microbial growth during storage, transport and retail.

That matters in a region where supply chains are often fragmented and temperature control can be inconsistent outside premium channels. In markets such as Vietnam, Indonesia and the Philippines, produce may move through multiple intermediaries before reaching consumers. Even in Singapore, where retail standards are high, the country’s reliance on food imports makes shelf life a national resilience issue as much as a supermarket concern.

Also Read: The environmental ethics of AI should be a product decision, not a sustainability footnote

Food waste is also a stubborn problem for Singapore. The city-state generates hundreds of thousands of tonnes of food waste each year, according to official waste statistics, while importing more than 90 per cent of its food. Any technology that can keep produce usable for longer without adding heavy infrastructure could therefore have relevance beyond a niche sustainability story.

Commercial traction beyond the lab

N&E is not entering this round as a company still searching for its first use case. Its customers already include Singapore Airlines and supermarket chain Sheng Siong. Its KeepWell hand sanitiser was also selected for inclusion in the official 2026 Singapore National Day Parade fun pack, putting the company’s technology in front of a mass local audience.

The new capital will be used to expand ViKANG99 across three areas: active antimicrobial food packaging, plant-based cleaning, and post-harvest agriculture. N&E also plans to support regulatory programmes, international market entry and team expansion.

Australia is shaping up as one of its key growth markets. Kinoya was appointed exclusive Australian distributor for the KeepWell range in June, with selected IGA supermarkets in Sydney lined up as an initial retail channel. KeepWell products are already being trialled by hospitality businesses including M Bar Thai Eatery, Show Nom Dessert and Show Neua Thai.

The Orange Wrap has also entered Australia’s foodservice market through Perth-based distributor Familiar Goods, which supplies hotels, restaurants and other foodservice operators.

For N&E, Australia offers a useful test bed. It is a sophisticated retail and foodservice market with strict food safety expectations, but it also faces long supply routes and high spoilage costs across fresh produce. Success there could strengthen N&E’s case with partners in other developed markets, including the UK and Germany, while giving it credibility in regional export markets such as Vietnam and Taiwan.

A crowded but growing shelf-life race

N&E is operating in a global market where several startups and established players are trying to extend food shelf life without relying solely on cold-chain expansion or traditional preservatives.

US-based Apeel Sciences, one of the best-known companies in the space, developed plant-derived coatings that slow water loss and oxidation in produce. Hazel Technologies, also from the US, works on packaging inserts that regulate ripening and reduce spoilage.

AgroFresh, a more established post-harvest technology player, provides freshness solutions for fruit supply chains, while startups such as Mori and Sufresca are exploring edible or biodegradable coatings.

N&E’s differentiation lies in its use of upcycled agricultural waste as an antimicrobial source and its attempt to apply the same active ingredient across packaging, cleaning and post-harvest treatment. That breadth could open several revenue channels, though it also means the company must navigate different regulatory regimes, customer buying cycles and product-performance expectations at once.

The company’s model is described as licensing-led, with ViKANG99 embedded into partner products rather than only sold through N&E’s own branded lines. If executed well, that could help it scale without building a large manufacturing footprint in every market. But licensing in food and packaging is rarely quick; partners typically require evidence on safety, performance, cost, durability and consumer acceptance.

Why investors are paying attention

Tundra Capital Managing Partner Timothy Hui said the investment fits the firm’s focus on technologies that address structural challenges in food systems.

“N&E Innovations is one of those startups that ticks every box; their technology transforms agricultural waste into a valuable ingredient, their antimicrobial active ingredient is applicable at various points across the food value chain, and they have demonstrated customers globally love their product,” Hui said.

The round also reflects a broader investor interest in climate-adjacent technologies that can show near-term commercial value. Food waste reduction has often been framed as an environmental goal, but for supermarkets, restaurants and growers, the argument is simpler: less spoilage means better margins.

Also Read: Why Southeast Asia’s next climate unicorn might be built from farm waste

For Southeast Asian startups, that commercial framing is important. Sustainability products can struggle when they are priced as a moral choice. They stand a better chance when they solve a cost, compliance or operational problem. N&E’s challenge now is to prove that its food-waste-derived antimicrobial can do so consistently across geographies and applications.

The company’s story is also unusually circular: agricultural waste becomes an ingredient that helps prevent more food from becoming waste. If N&E can translate that logic into scalable products and partnerships, it could turn a Singapore lab-born idea into a practical tool for food systems well beyond the region.

The post N&E Innovations nets US$1.7M to turn cashew husk waste into fresh produce protection appeared first on e27.

For decades, talent management has operated on a seemingly logical premise: identify your top performers, your A-players, and invest in them disproportionately. This approach, popularised by McKinsey’s War for Talent in the late 1990s, promised that organisations could secure competitive advantage by systematically differentiating their workforce. The logic was seductively simple: measure performance annually, rank employees on a curve, and focus resources on the highest-rated segment.

The universal A-player framework spread across industries because it offered HR leaders a simple, comparable metric for talent quality. If every role could be measured against the same scale, then talent could be managed like capital, allocated efficiently, tracked rigorously, and optimised continuously. This aspiration reached its logical conclusion in the talent supply chain movement, where frameworks like the R7 model (Recruit, Rate, Retain, Redeploy, Redevelop, Release, Rehire) attempted to apply manufacturing discipline to workforce management.

The problem is that the universal A-player never actually existed. Consider a global bank that rated a call centre agent and an investment banker on the same 1-5 scale. The agent who resolved 95 per cent of calls on first contact received a 3 because her manager was a tough rater. The banker who closed a mediocre deal received a 5 because his manager was lenient. The bank then used these ratings to allocate training budgets and promotion opportunities. Research has found that individual rater biases can account for up to three times as much variance in performance scores as actual employee performance. When a manager rates an employee, the result reveals more about the manager’s evaluation style than about the employee’s capabilities.

This bias problem becomes exponentially more dangerous when flawed historical ratings are used as training data for AI systems. Consider Amazon’s AI recruiting tool: trained on resumes submitted over a decade, it learned to penalise resumes containing the word women’s (e.g., Captain of women’s chess club) because most past hires were men. Algorithms trained on biased human decisions do not eliminate bias. They scale it with terrifying efficiency.

Why the universal A-player cannot survive the AI-driven supply chain

The talent supply chain metaphor demands real-time visibility into quality, throughput, and defects. Yet the universal A-player provides none of these. It suffers from three terminal pathologies.

First, it is static. Annual performance reviews look backward six to twelve months. Consider a cybersecurity firm whose annual reviews in December rated engineers on skills relevant to pre-AI threat landscapes. By January, generative AI had transformed the work entirely. Those A-players from December lacked the capabilities needed for February’s critical incident response. The organisation knew who was excellent yesterday, not who was right for today.

Second, it is retrospective. The universal A-player relies entirely on lagging indicators, completed projects, closed deals, past ratings. Consider a retail chain during the pandemic shift to e-commerce. Their A-player store managers, rated highly on in-person sales, struggled to adapt to digital fulfilment. Meanwhile, a previously B-player assistant manager who had built the store’s social media presence in her spare time turned out to be the ideal candidate. The lagging indicators missed her entirely.

Third, it is universal. Traditional A-player frameworks apply generic traits, leadership, initiative, strategic thinking, across wildly different roles. Consider a manufacturing company that rated a warehouse supervisor and a data scientist on the same strategic thinking competency. The supervisor, whose role demanded real-time operational decisions about shift scheduling, scored poorly. The data scientist, who spent weeks developing long-term forecasting models, scored highly. The company concluded the supervisor was a B-player and redirected development funds. Six months later, the supervisor had reduced warehouse defects by 40 per cent through a simple reorganisation. The universal scale had measured the wrong thing.

Also Read: Are you a human resource?

The new model: Dynamic talent matching for the AI era

The end of the universal A-player opens the door to dynamic talent matching based on real-time, role-relative fit. This new model solves both the measurement problem and the bias problem.

• Pillar one: From static to dynamic

Instead of annual ratings, the new model measures talent continuously against specific role demands. A global consumer goods company transformed its recruitment process by abandoning CVs and annual reviews entirely, replacing them with game-based assessments and video interviews analysed by AI. Candidates completed neuroscience-based games measuring problem-solving and learning agility in under 25 minutes, with results evaluated against role-specific benchmarks, not universal scales. The approach increased new hire diversity while reducing screening time by 75 per cent.

Research from Heliyon demonstrates that skill assessments can be recalibrated using neural networks that combine expert knowledge with real-time data, ensuring measurement adapts as roles evolve.

• Pillar two: From retrospective to predictive

The new model uses predictive signals instead of lagging indicators. In one documented implementation, candidates completed open-ended questions like How do you know you’ve understood what someone said? without seeing skill labels. The AI mapped free-text responses to the five to ten core tasks of the specific role. This approach prevented self-screening bias, candidates who doubted their qualifications applied anyway, and eliminated the correct answer bias of multiple-choice tests. The result: retention increased by 25 per cent and time-to-hire dropped by 23 per cent.

Research from Taylor & Francis on adaptive job recommendation systems demonstrates that such approaches achieve 92 per cent accuracy in job matching while reducing algorithmic bias by 15 per cent through adversarial debiasing mechanisms.

• Pillar three: From universal to role-relative

The new model recognises that different critical roles require different measurement criteria. A manufacturing company implemented role-specific validation by studying 226 employees, correlating assessment scores with actual job outcomes. For warehouse associates, top scorers performed at the 64th percentile on the job, while bottom scorers performed at the 33rd percentile. Top scorers were also three times less likely to be involved in safety incidents. For data analysts in the same company, the predictive criteria were completely different: pattern recognition speed and intellectual curiosity, not physical safety indicators.

Also Reda: The app worked, the product didn’t: Can we install judgement into AI agents?

Solving the bias problem through dynamic matching

The universal A-player made bias invisible. Dynamic talent matching makes bias detectable and correctable through three mechanisms.

First, role-relative measurement breaks proxy discrimination. Consider a health system where universal ratings penalised nurses who took family leave, marking them as lower commitment. Under role-relative measurement, a nurse’s fit for an intensive care role depends on clinical judgement and rapid response time, not attendance patterns from three years ago. The proxy of leave-taking no longer leaks into the prediction.

Second, continuous auditing replaces one-time validation. A technology company’s engineering team documented that the most effective bias mitigation is continuous fairness auditing using the EEOC’s four-fifths rule: a model’s scoring rate for any demographic group must be at least 80 per cent of the highest-scoring group’s rate. Leading implementations audit intersectional groups, Black women, Latinx men, where bias is often most severe. When a financial services firm ran these audits, they discovered their model systematically downgraded candidates from historically Black colleges. They retrained the model, removing the proxy signal, and hiring diversity improved without reducing performance.

Third, human-in-the-loop governance ensures accountability. Under the EU AI Act, hiring algorithms are classified as high-risk systems requiring full risk-management programmes. Employees and candidates must have the right to understand what signals influence fit scores and to contest inferences. Consider a logistics company where an algorithm flagged a driver for low reliability based on GPS data showing frequent stops. The driver contested, explaining the stops were required safety checks for hazardous materials. The system was adjusted, and the driver became one of the highest-performing team members.

Conclusion: The universal A-player is dead

The universal A-player was a useful simplification for a stable world. That world no longer exists. In an AI-driven talent supply chain where roles evolve continuously and critical talent must be redeployed in days, static, retrospective, universal ratings are worse than useless.

The new model asks a fundamentally different question: not who are our A-players? but who fits this role, right now? By shifting from static to dynamic measurement, from retrospective to predictive signals, and from universal to role-relative standards, organisations can finally escape the bias trap and manage talent with real-time discipline.

Organisations that cling to the universal A-player will find themselves measuring what no longer matters. Those that embrace dynamic talent matching will redeploy capability as quickly as they respond to changing demand. The supply chain taught us to manage inventory in real time. It is time to manage talent the same way.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post The end of the universal a-player: Dynamic talent matching in the AI-driven supply chain appeared first on e27.

AI is remarkably good at producing answers. It is even better at sounding certain. Ask a difficult question and, within seconds, a system can gather information, connect ideas and return a polished explanation. Yet a harder question arrives later: what happens when reality proves the answer wrong?

I met that problem while building OnTheRice, a Singapore-based AI research publication. Part of its work involves making time-bound directional calls, saving the evidence available at the time, and checking the result later. The experience changed how I think about AI products. Producing an answer was the easy part. Preserving an honest record of it was much harder.

Freeze the answer before reality arrives

A prediction is easy to admire while its outcome is unknown. It is also easy to repair after the fact. A changed sentence, a missing timestamp or a quietly removed failure can turn poor judgement into a convincing success story.

We began locking each call before its result was known: the direction, entry price, source set, publication time and evaluation time. As at 1 September 2026, the public Founder Ledger showed 52 correct and 34 incorrect results across 86 resolved calls, or 60.47 per cent. 10 older records remained visible but excluded because they could not be verified properly.

Those figures are not proof that the system is exceptional. They are useful because they are incomplete, imperfect and inspectable. Once the misses remained on the page, wrong stopped being one category. Sometimes the reasoning failed. Sometimes the information arrived too late. Sometimes the market had already moved. Sometimes an event changed the conditions after publication.

The practical lesson is simple: save the original output, the evidence, the timestamp and the scoring rule before the outcome arrives. Otherwise, learning can become hindsight wearing a lab coat.

More links do not mean more evidence

A second problem appeared when the system began reading large numbers of reports. 10 websites may cover the same event, yet nine might trace back to one wire story. That is not ten witnesses. It is one witness with excellent distribution.

AI systems can confuse information volume with independent confirmation. Counting URLs rewards duplication. It can make a thin claim look strong simply because it travelled far.

Also Read: SEA’s AI boom has a water problem it cannot offset away

We started treating source origin as part of the evidence. Reports were grouped when they repeated the same underlying account, while genuinely independent reporting carried more weight. The practical rule is to trace claims backwards, not merely count how many pages repeat them. Three independent reports can tell us more than 100 copies.

Time belongs inside the evidence

Suppose a report at 8am says oil is falling and another at 5pm says it is rising. Which is wrong? Possibly neither. The world happened between them.

An AI system that ignores time can flatten both statements into one moment. Worse, it can use information published later to explain a decision made earlier. This creates the illusion that the system knew more than it could have known.

Useful records therefore need more than a source link. They need the time the source was published, the time the system found it, the time the output was made and the time it was assessed. This applies beyond markets. A business cannot honestly link a competitor’s price change to falling sales without knowing which came first.

Documentation is not decoration

This approach is not unique to one product. The researchers behind Model Cards for Model Reporting proposed standardised records of a model’s intended uses, evaluation and limitations. The NIST AI Risk Management Framework also treats accountability, transparency, validity and reliability as central parts of trustworthy AI.

The need is growing. Stanford’s 2025 AI Index reported 233 AI-related incidents in 2024, 56.4 per cent more than in 2023. The figure does not mean every incident could have been prevented by better records. It does show why explanations offered only after something goes wrong are not enough.

For an applied AI product, a small receipt can carry the source, timestamp, system version, original output, confidence level, known limits, evaluation rule and eventual outcome. None of this looks as exciting as a smarter model demonstration. It is far more useful during a dispute, audit or failure review.

Also Read: The app worked, the product didn’t: Can we install judgement into AI agents?

Let the system say I don’t know

AI products are designed to answer. Silence looks broken, especially in a demonstration. But forcing a decision when sources conflict or evidence is missing creates artificial certainty.

One of our hardest lessons was to separate wrong from unverifiable. The first means reality contradicted a recorded call. The second means the evidence is too weak to score it honestly. Combining them hides different problems; counting either as a win is worse.

A mature system needs permission to abstain. I don’t know yet should be a valid output when confidence falls below a clear threshold. Teams should record why the system abstained, then test whether the rule was too cautious or appropriately restrained.

Trust needs evidence

The AI industry is understandably focused on better reasoning, larger context windows and stronger models. Yet applied systems face a less glamorous test: can another person inspect what happened?

Can they see the original source? Can they see what the system said, when it said it and what changed afterwards? Can they find the failures as easily as the successes?

Building an AI that keeps receipts taught me that mistakes are not embarrassing leftovers. Properly preserved, they are training data for the product team, evidence for the user and a guard against self-deception.

A system should not earn trust by describing itself as intelligent. Show the work. Keep the misses. Let the record speak.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post I built an AI that keeps receipts. The mistakes became the useful part appeared first on e27.

Every founder has opinions.

After all, they spend their days solving customer problems, navigating market uncertainty, raising capital, hiring talent and making decisions that shape the future of their business. Those experiences naturally produce perspectives on everything from emerging technologies and regulation to leadership, innovation and industry trends.

Yet expertise alone does not earn media coverage.

Every week, journalists receive pitches from founders who want to share their views on the latest developments in their industry. Most are ignored, not because the founders lack credibility, but because their perspectives have not been translated into stories that are timely, relevant or valuable to readers.

That is the difference between having an opinion and having a media-ready narrative.

The role of a startup PR agency or tech PR agency extends well beyond writing opinion pieces or arranging interviews. Effective thought leadership begins with understanding how journalists think, what audiences care about and how a founder’s expertise can contribute meaningfully to conversations already taking place.

When those elements come together, founder insights become more than commentary. They become stories that build credibility and strengthen a company’s reputation.

Start with the media landscape, not the founder

Many companies begin by asking what their founder wants to say.

A better question is where that conversation belongs.

Every publication serves a different audience and has its own editorial priorities. Business publications may look for commentary on market trends, while technology media often seeks practical insights into innovation, product development or investment activity. The same founder perspective can be highly relevant to one publication and completely unsuitable for another.

This becomes even more important for companies operating across Southeast Asia.

Although the region is often discussed as a single market, each country has its own media ecosystem, business priorities and cultural context. A narrative that resonates with journalists in Singapore may require a different angle in Indonesia, Malaysia or the Philippines.

For example, a founder discussing artificial intelligence could focus on regulatory frameworks for Singaporean business media, manufacturing transformation in Vietnam or digital inclusion in Indonesia. The expertise remains consistent, but the story evolves to reflect local priorities.

The best thought leadership is never one-size-fits-all. It respects regional nuances while remaining authentic to the founder’s perspective.

This is why an experienced corporate communications agency develops narratives with both editorial relevance and local market context in mind.

Understand what journalists are trying to write

Successful media relations is not simply about understanding publications. It is about understanding the journalists behind them.

Every reporter develops areas of expertise and recurring themes throughout their career. Some focus on breaking news, while others specialise in long-form analysis, policy developments or emerging technologies. Many spend years covering the same industries and are constantly looking for credible voices who can help explain what is happening beneath the headlines.

Also Read: The role of thought leadership in scaling beyond your first market

The strongest founder commentary supports those objectives.

Rather than asking, “What do we want to say?”, communications teams should ask, “What questions is this journalist already trying to answer?”

Can your founder explain why investment in a particular sector is accelerating? Have they observed changing customer behaviour before industry reports identified the trend? Can they provide practical context behind a new government policy or technological development?

When founder expertise helps journalists tell a better story, media opportunities become significantly easier to secure.

This is one reason companies work with a PR agency in Singapore. The media pool tends to be smaller in comparison, and strong media relations help you understand editorial priorities and shape founder insights to create genuine value for journalists and their audiences.

Genuine expertise is more valuable than manufactured opinions

There is a common misconception that thought leadership requires controversial opinions or bold predictions.

In reality, credibility consistently outperforms sensationalism.

Founders possess something that analysts and commentators often cannot replicate: first-hand operational experience. They understand customer challenges, market dynamics and competitive pressures because they navigate them every day.

Those experiences produce insights that are both practical and distinctive.

Instead of encouraging founders to manufacture provocative opinions, communications teams should identify recurring patterns, lessons and observations that have emerged through building the business.

Sometimes the most compelling narrative is not about predicting the future. It is about helping people understand the present.

Journalists value perspectives grounded in experience because they offer readers something genuinely useful rather than simply adding another opinion to an already crowded conversation.

Every narrative should strengthen your company’s positioning

Not every interesting opinion deserves to become thought leadership.

The most effective founder narratives reinforce the expertise that a company wants to become known for.

A cybersecurity company should consistently contribute to discussions around digital resilience and cyber risk. A fintech founder should become recognised for perspectives on financial innovation, regulation and customer trust. A climate technology business should build authority around sustainability and industrial transformation.

Also Read: Choosing the right tool: What works for news, thought leadership, influence

Over time, these repeated associations shape how journalists, customers, investors and industry peers perceive the organisation.

Visibility alone is rarely the objective.

Authority is.

Every interview, contributed article and expert quote should strengthen the same strategic positioning. This is where an experienced tech PR agency provides lasting value, ensuring media opportunities contribute to long-term reputation rather than short-term exposure.

Timing is often the deciding factor

Even exceptional insights can fail if they arrive at the wrong moment.

Media operates on relevance. Journalists are constantly looking for expert perspectives that help explain developments already shaping the news cycle, whether that involves funding announcements, new regulations, economic uncertainty, emerging technologies or changing customer behaviour.

The most successful communications strategies therefore connect founder expertise with conversations that are already happening.

Instead of waiting for inspiration, communications teams should actively monitor industry developments and identify opportunities where their founder can contribute a credible, differentiated perspective.

When expertise meets timing, media-ready narratives become significantly more compelling.

Turning expertise into influence

Founders rarely struggle with having opinions. More often, they struggle with communicating those opinions in ways that resonate beyond their own organisation.

Media-ready narratives are built at the intersection of expertise, editorial relevance, strategic positioning and timing. They are grounded in genuine experience, shaped around the needs of journalists and aligned with the company’s long-term business objectives.

That is why effective thought leadership is never simply about publishing more content. It is about consistently contributing meaningful perspectives that help audiences better understand their industry.

Whether you’re an emerging startup preparing for your first round of media engagement or an established technology company expanding across Southeast Asia, the goal remains the same.

Transform founder expertise into stories that journalists want to tell, audiences want to read and stakeholders remember long after the headlines have disappeared.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post How to turn your founder’s opinions into media-ready narratives appeared first on e27.

It was not.

Our HR team wanted employees to share more about company culture, hiring updates, internal events, and project wins on LinkedIn. We were not trying to turn everyone into corporate influencers. We only wanted people to feel more comfortable writing about work when there was something worth sharing.

The usual problem was the blank page. Someone attended a company event, joined a volunteer activity, completed a project, or had a good team moment. They were willing to post, but they did not know how to start. Some people worried about sounding too formal. Some worried about saying the wrong thing. Some simply did not have time to turn a rough thought into a clean post.

So we set up a ChatGPT Project for it.

The idea was simple. Instead of every employee prompting from scratch, HR would condition the project with shared context: company tone, basic social media guidelines, examples of acceptable employee posts, phrases to avoid, and reminders about confidential information. Anyone using that project would start from the same set of instructions.

At first, it worked well.

Employees could paste in a few rough notes and ask for a LinkedIn draft. Someone could write, “We had a team volunteering event last Friday, I helped with registration, make this sound natural.” Another person could ask for a short post about joining the company, a project milestone, or a hiring announcement. The drafts came out cleaner than what most people would write from scratch.

For HR, this looked like progress. More people were willing to share. The posts were more readable. The tone was more consistent. Employees were not staring at a blank page anymore.

Also Read: Quantum’s ChatGPT moment is coming — and it’s worth trillions

Then we noticed the issue. Some drafts sounded a little too official.

An employee might only want to say they enjoyed a team event, but the draft would turn it into a polished statement about company values. A hiring post might sound like it came from the corporate page instead of a person. A project post might imply that the employee was speaking on behalf of the company, even when they were only sharing a personal reflection.

That was the friction point. The ChatGPT Project had made drafting easier, but it had also made company-sounding posts easier to produce.

This mattered because employee advocacy sits in an awkward middle ground. A personal LinkedIn post is not the same as an official company announcement. But when the post is about work, people can easily blur the line. A sentence that sounds harmless inside a draft can still create confusion outside the company.

The issue was not that the project was bad. The shared context helped. It reduced blank-page writing and gave employees a safer starting point. The problem was that our instructions focused too much on writing style and not enough on posting boundaries.

So we changed the project instructions.

We added a simple rule: before drafting, ChatGPT had to identify what kind of post it was preparing. Was it a personal reflection, a hiring-related post, an event recap, a project update, or a repost of official company news?

That small step changed the output.

Also Read: OpenAI’s Astra aims to turn AI from chatbot into digital worker

If the post was a personal reflection, the draft had to sound like the employee. It could mention what they experienced, what they appreciated, and what they learned, but it should not make broad claims on behalf of the company. If the draft crossed that line, ChatGPT had to flag it to the user.

If the post was about hiring, it had to stay close to the approved hiring message. It could make the post warmer or more personal, but it should not invent role details, benefits, salary hints, or promises about career progression.

If the post was about a project, it had to avoid confidential details, client names, unreleased work, internal numbers, or claims that had not already been cleared.

We also asked the project to show a short “check before posting” note under each draft. Not a long policy lecture. Just a few practical checks: remove confidential details, confirm that this is your personal view, check whether the role or event information is current, and get approval if the post sounds like an official company statement and not personal.

That helped because employees did not need to remember the whole social media policy every time. The reminder appeared where the work happened.

It also showed us the limitation of using a ChatGPT Project alone. The project could draft and warn, but it could not route approvals by itself. Right now, we are moving beyond the project setup so the workflow can become more agentic. When approval is needed, the workflow should route the draft to the right person, such as Marcom, Legal, or HR. That person can adjust it, send it back to the writer and it will automatically be posted.

This was more useful than writing a long rulebook. The workflow became less about generating perfect posts and more about helping employees stay inside the right lane.

We did not need employees to become marketing people. We needed them to have a safer way to share real work moments without accidentally sounding like the corporate communications team.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post I used to think the hard part of employee advocacy was getting people to post appeared first on e27.

The next customer for a software company may not be a person scrolling through a pricing page. It may be an AI coding agent, asked by a developer to choose a database, install an SDK, connect an API, or fix a broken integration.

That shift is still early, but it is already changing how software is discovered and adopted. Tools such as Claude Code, Codex, Cursor, GitHub Copilot and OpenCode can now search for products, compare options, read documentation and write implementation code on behalf of users. In that world, a company’s website is no longer the only front door. Its documentation, APIs, SDKs, command-line tools and machine-readable interfaces become part of the sales funnel.

Also Read: The app worked, the product didn’t: Can we install judgement into AI agents?

Lightsage, a San Francisco-based startup founded by Singaporean founders Jun Liang Lee (CEO) and Sean Er (CTO), wants to build the analytics layer for this new behaviour. The company has raised US$4 million in funding led by Nexus Venture Partners to develop what it calls an Agent-Led Growth platform.

The round also includes operators from the developer tools and AI ecosystem, including former Salesforce CTO Steven Tamm, Postman CEO Abhinav Asthana, Apollo CEO Matt Curl, DocuSign President and GM of Growth Robert Chatwani, Resend CEO Zeno, Firecrawl co-founder Eric, Daytona CEO Ivan, Tinyfish COO Shuhao, and others.

From product-led to agent-led growth

For the past decade, many software companies have optimised around product-led growth, a model in which users discover, try and pay for products with limited involvement from sales teams. The playbook was built around human behaviour: search rankings, landing pages, onboarding flows, product analytics, emails and conversion funnels.

AI agents disrupt that pattern because they may compress discovery, evaluation and implementation into a single workflow. A developer might ask an agent to “add payments to this app” or “find the best OCR API for invoices”, and the agent could decide which vendor to use before the human ever sees a shortlist.

That matters for Southeast Asia, where many startups run lean engineering teams and rely heavily on global developer infrastructure. A fintech in Jakarta, a logistics startup in Ho Chi Minh City or a SaaS company in Singapore may increasingly use AI coding tools to speed up product work. If those agents default to familiar global vendors, newer or regional software companies could struggle to be discovered, even if their products are technically strong.

Lightsage’s core argument is that visibility in AI-generated answers is only one part of the problem. A product may be mentioned by an AI assistant, but still lose the “agent customer” if its documentation is confusing, its SDK fails, its authentication flow is unclear, or its API examples cannot be executed reliably.

“We are moving from an internet where AI tells people which software to use to one where AI increasingly uses the software itself,” said Lee. “Visibility still matters, but the real test is whether an agent can understand your product and get to a successful outcome.”

How the platform works

Lightsage allows software companies to see their product through an agent’s perspective. Its platform runs simulations across answer engines and coding agents, checking not only whether a company appears against competitors, but what happens after that.

Agents are assigned real tasks that require them to navigate documentation, pick the right tools and use APIs, SDKs, CLIs, Model Context Protocol servers and agent skills. MCP, an emerging standard popularised in the AI developer ecosystem, allows AI systems to connect more easily with external tools and data sources.

Also Read: AI agents could help Southeast Asian firms untangle cross-border payment costs

When the agent fails, Lightsage identifies where the workflow broke. The issue could be poor discoverability, missing examples, inconsistent documentation, authentication problems, an API endpoint, an SDK implementation, or an incompatible MCP server. Teams can then fix the issue, rerun the test and measure whether the agent completes the task more successfully.

The company also tracks real agent traffic, showing when agents visit a customer’s website or documentation, what they interact with, and whether those journeys lead to product usage. Over time, Lightsage wants to feed those insights back into development and deployment workflows so products can continuously improve for both agents and humans.

The platform currently supports Claude Code, Codex, Cursor, GitHub Copilot, OpenCode and other coding agents.

Early users and a new analytics gap

Lightsage is beginning with developer software, where the behaviour of coding agents is easiest to observe. Its early customers include Firecrawl, Reducto, Daytona, Rime and Tinyfish. These companies use the platform to understand why agents recommend certain products, where integrations fail and whether changes to documentation or product flows improve outcomes.

A typical case may start with a company discovering that a coding agent keeps recommending a rival. Lightsage then recreates the same task across multiple products and agents to determine whether the problem is awareness, documentation quality, or the actual product experience.

This is an analytics gap that traditional software tools were not designed to handle. Human acquisition is usually tracked through search terms, ad clicks, referral links, sign-ups and sales calls. Agents may not follow those paths. They can discover a product through generated answers, inspect documentation directly, call an API and influence a purchase without producing the same trail of clicks.

Their behaviour is also unstable. Different coding agents can approach the same task differently, and their preferences may shift as models are updated. A workflow that works for one agent may fail in another.

Abhishek Sharma, partner at Nexus, framed this as a wider change in online commerce. “AI is now shifting that agency from humans to agents, which can discover, evaluate and act on a customer’s behalf,” he said. “Lightsage is building the intelligence infrastructure for this new era of autonomous browsing, helping companies optimise for agent conversion, not just awareness.”

Competitive landscape

Lightsage sits at the intersection of AI search optimisation, developer experience testing and product analytics. Its closest rivals are likely to come from several directions rather than one neat category. Generative engine optimisation startups such as Profound, AthenaHQ, Scrunch AI and Peec AI help brands understand how they appear in AI answers, while developer observability and AI infrastructure tools such as LangSmith, Helicone and Langfuse focus on monitoring AI applications and model behaviour.

Traditional product analytics companies, including Amplitude, Mixpanel and PostHog, already help software teams understand human users. Lightsage is betting that agent behaviour will become distinct enough to need its own system of record.

The challenge is whether “agent-led growth” becomes a durable software category or remains a feature added by existing analytics and developer tools platforms. Large incumbents already own parts of the workflow, from code assistants to API platforms and observability stacks. Lightsage will have to show that agent discovery, agent experience and agent attribution are not just interesting signals, but commercial levers that affect revenue.

Why it matters for Southeast Asia

For Southeast Asian startups, the rise of agent-led software adoption cuts both ways. On one hand, small teams can use AI agents to build faster, integrate complex tools and compete more effectively with better-funded rivals. On the other, if agents concentrate attention on a narrow set of familiar vendors, local or emerging software companies may find it harder to break into global workflows.

This could be especially relevant in areas where the region is producing more infrastructure and B2B software, from fintech APIs and compliance tools to logistics software and vertical SaaS. Winning a human developer’s trust may no longer be enough. Products will also need to be legible to machines.

Also Read: When AI agents start acting on our behalf, security gets more complicated

Lightsage plans to use the new funding to expand its agent evaluation, analytics, attribution and optimisation capabilities across APIs, SDKs, CLIs, MCP servers and agent skills. Developer tools are the starting point, but the company expects agent behaviour to spread into B2B software, infrastructure and payments.

If that happens, the old growth funnel may not disappear, but it will have a new participant. The buyer may still be human. The first user may increasingly be an agent.

The post Singaporean founders’ Lightsage bags US$4M to decode how AI agents choose software appeared first on e27.

The Philippines’s electric vehicle (EV) market is expected to keep growing in 2026 even as the broader car industry contracts, underscoring how fuel costs, government incentives and cheaper Chinese models are beginning to shift buyer behaviour in one of Southeast Asia’s more difficult markets for electrification.

BMI Country Risk and Industry Research forecasts electric vehicle sales in the Philippines to rise 11.2 per cent year-on-year in 2026, from 29,479 units to 32,776 units. Its definition of EVs includes battery electric vehicles and plug-in hybrids, but excludes conventional hybrids that cannot be charged externally.

Also Read: Southeast Asia’s EV startups draw US$622M as clean mobility shifts from pitch to pilot

That growth is notable because BMI expects total vehicle sales in the country to fall 8.7 per cent to 423,750 units next year. As a result, EV penetration is projected to climb from 6.4 per cent in 2025 to 7.7 per cent in 2026.

The numbers point to a familiar pattern across Southeast Asia: EV adoption is rising, but not evenly. In wealthier urban centres, buyers are increasingly open to electrified vehicles, particularly as fuel prices climb and more models enter showrooms. Outside those areas, charging access, road conditions and affordability continue to slow the transition.

Fuel prices cut both ways

BMI said higher fuel prices linked to the US-Iran conflict are producing two opposing effects in the Philippine auto market. On one hand, they are weakening overall demand by raising transport and ownership costs. On the other, they make electrified vehicles more attractive to buyers who can still afford a new car.

“Consumers able to purchase a new vehicle have a stronger incentive to consider models that can reduce fuel expenditure. This will benefit BEVs, PHEVs and HEVs, particularly among higher-mileage drivers and urban consumers,” BMI said.

That matters in the Philippines, where traffic congestion, long commute times and frequent stop-start driving can make fuel spending a major concern for households and small businesses. For fleet operators, ride-hailing drivers and urban professionals, the economics of electrification can become more compelling when petrol prices are volatile.

Still, the shift is not a simple move from petrol cars to fully electric ones. BMI expects hybrids and plug-in hybrids to remain important because they offer lower fuel use without forcing drivers to depend fully on public charging networks. This is especially relevant in a market where many consumers do not have reliable access to home charging.

Chinese brands push prices lower

A broader model line-up is also helping the market. BMI pointed to brands such as BYD, Chery, MG, VinFast and Tesla as supporting the local battery EV segment through more competitive pricing and a wider choice of vehicles.

Chinese automakers are likely to be particularly important. BYD, MG, GAC Aion and Chery are bringing more affordable electric SUVs and crossovers into the market, which could help reduce the price gap between EVs and comparable internal combustion engine vehicles.

“We contend that the expansion of Chinese brands will be particularly important because it will reduce the price gap between EVs and comparable internal combustion engine vehicles while increasing consumer awareness of electrified technology,” BMI said.

This mirrors developments elsewhere in Southeast Asia. Chinese EV makers have moved aggressively into Thailand, Indonesia and Malaysia, often using price, financing and high-spec models to challenge Japanese incumbents. The Philippines has been slower to electrify at scale, but the entry of more Chinese brands could alter expectations among buyers who previously saw EVs as niche or luxury products.

Also Read: Datakrew’s US$2.6M raise is a bet on the EV problem nobody wants to own: battery failures

The popularity of SUVs, crossovers and pickup trucks adds another layer. These are important segments in the Philippine market, but electric versions tend to carry higher sticker prices and require larger batteries. More competitively priced Chinese models could therefore play a decisive role in pushing EVs beyond early adopters.

Policy helps, but charging remains the bottleneck

Government policy is another tailwind. Battery EVs are exempt from import tariffs and excise taxes until 2028, while hybrids and plug-in hybrids benefit from tariff exemptions and lower excise tax rates. These measures are narrowing the price gap with petrol and diesel vehicles.

The Electric Vehicle Industry Development Act also provides a policy framework for charging infrastructure, regulatory support and the use of EVs in public-sector fleets. It sets a target for EVs to make up at least 10 per cent of government vehicle fleets, a move that could create demand while making the technology more visible to the public.

But charging infrastructure remains the clearest obstacle. The Philippines currently has around 1,600 charging points, according to the Electric Vehicle Association of the Philippines. These include 781 alternating current chargers, 291 direct current chargers and 528 battery-swapping stations.

“Although the network is expanding, charging facilities remain concentrated in major urban areas,” BMI said.

That concentration limits the practicality of battery EV ownership for people without home charging and for drivers who travel between cities or through less-developed areas. The dominance of alternating current chargers also means charging can be slow, making plug-in hybrids a more comfortable compromise for many consumers.

Range anxiety is not only about how many chargers exist. Buyers also need confidence that chargers are in the right places, working when needed, compatible with their vehicle and not already occupied. In a market with uneven road quality and heavy congestion, those concerns become more pronounced.

This is why the Philippine market may not follow the same path as countries that built dense charging networks early. Instead, electrification could advance through a mix of battery EVs in cities, plug-in hybrids for flexibility and conventional hybrids for buyers who want lower fuel costs without changing refuelling habits.

A stronger second half of the decade

BMI expects EV sales growth to accelerate after 2026. It forecasts sales to reach 51,666 units in 2027, 65,432 units in 2028, 78,781 units in 2029 and 91,730 units in 2030. That implies average annual growth of 29.3 per cent between 2026 and 2030.

EV penetration is forecast to rise from 7.7 per cent in 2026 to 11.4 per cent in 2027, 13.7 per cent in 2028, 15.6 per cent in 2029 and 17.3 per cent in 2030. Internal combustion engine vehicle sales, by contrast, are expected to recover more slowly, from 390,974 units in 2026 to 438,106 units in 2030.

Also Read: Inside Thailand’s EV and battery push: Balancing growth with sustainability

The outlook suggests electrified vehicles will account for a growing share of industry expansion in the second half of the decade. But two risks stand out: slow charging infrastructure rollout and possible changes to incentives after 2028. Either could weaken affordability and consumer confidence.

The upside case is equally clear. If Chinese automakers intensify competition, battery costs continue to fall and public and private investment in charging accelerates, EV adoption could move faster than BMI currently expects. For the Philippines, the question is no longer whether electrification will happen, but whether infrastructure and policy can keep pace with the market’s growing interest.

The post EVs gain ground in the Philippines as fuel costs squeeze drivers appeared first on e27.

The crypto market is working through a sharp pullback. In the last 24 hours, total market cap fell 1.74 per cent to US$2.62T. Bitcoin led the decline. The largest digital asset dropped 1.87 per cent to US$76,796.54. This move did not happen alone. It tracked a broader sell-off across traditional finance. The 91 per cent correlation with the S&P 500 shows macro forces are driving price action. That link matters because crypto is trading as a risk asset, not as a safe haven.

The US PPI for August arrived on September 10 and served as the main trigger. Annual inflation reached 5.4 per cent. That result came in above expectations. The data revived fears that the Federal Reserve will raise rates or keep policy tight. Treasury yields moved higher. Stocks, gold, and crypto all fell at the same time. Bitcoin sold off as liquidity concerns and weaker investor appetite hit risk markets. The next test is the CPI report on September 11. That print will shape the odds of a rate hike before the Fed meeting on September 15-16.

Market structure made the drop worse. A wave of long liquidations hit the derivatives market. Over 24 hours, forced liquidations closed US$96.53M in BTC positions. Long bets made up 90 per cent of that total. Total open interest had risen 10.84 per cent before the move. New leverage had entered the system. When prices turned lower, forced selling created a feedback loop. That loop added downward pressure beyond what the inflation report alone would suggest. The same pattern appeared across the wider market. BTC liquidations totalled US$96.53M over 24 hours, up 16.35 per cent. Open interest across derivatives rose 9.11 per cent to US$462.13B. High leverage remains a clear risk.

Sentiment has cooled. The Fear & Greed Index fell from 75 last week to 67. That shift moved the mood from extreme greed to caution. The total market cap broke below its pivot point of US$2.61T. It also fell under its seven-day moving average of US$2.67T. The August rally has lost momentum. The sector is now in a corrective phase as speculative heat leaves the market. The RSI(7) sits at 30.02. That reading suggests the market is nearing oversold territory. If buyers return, the decline could slow.

Also Read: Will Bitcoin hold US$77,000 or drag the market to US$2.51T? The September 10 answer

For Bitcoin, the immediate test is the US$76,000 to US$76,350 range. This zone reflects the average cost basis for active investors. Buyers have defended it many times. If the price holds above this support, Bitcoin could stabilise and trade between US$76,000 and US$79,400 before the CPI release. A decisive break below US$76,000 would open a path toward US$73,000. A close above US$79,400 would signal a return to the uptrend.

For the total market, the path depends on the US$2.59T level. That level matches the 23.6 per cent Fibonacci support. A hold there could lead to range-bound trading between US$2.59T and US$2.67T. A daily close below US$2.59T would target the next support at US$2.51T, the 38.2 per cent Fibonacci level. A US$2.67T reclaim would indicate stabilisation. The market remains in a corrective phase within a larger uptrend. It is still up 19.65 per cent over 30 days. The key question is whether support attracts buyers or whether liquidations drive a deeper flush.

Macro data will likely decide the next move. The CPI release is the major cue. A hot reading could extend the risk-off mood and trigger another leg down. A cooler reading could give buyers a reason to step in. The correlation with the S&P 500 reached 91 per cent and 92 per cent in two readings. The link with gold reached 87 per cent. This highlights how sensitive crypto has become to macro liquidity sentiment. This is not a market-moving story on its own. It is moving with the broader financial system.

On the regulatory front, Senate Republicans released a revised CLARITY Act. The bill runs 630 pages. It aims to establish a federal framework for digital assets while tightening regulations on non-decentralised DeFi and stablecoins. Senator Cynthia Lummis led the effort. The text includes more than 100 changes requested by Democrats. It rebrands Title I as the Lummis-Gillibrand Responsible Financial Innovation Act of 2026. Lawmakers designed the bill as a broad federal framework rather than a narrow crypto adjustment.

The revised bill would split oversight between the SEC and CFTC. It would formalise how regulators would oversee exchanges, brokers, and protocols. It introduces detailed language for non-decentralised finance protocols that people or groups can control or materially alter. Those protocols would have to register with the CFTC. The text narrows DeFi rules to spot and cash digital commodities. For DeFi, the bill draws a line between genuinely decentralised protocols and those that are decentralised in name only. Participating in governance or security councils does not, by itself, constitute control. Protocols that identifiable actors can steer would be subject to CFTC registration and rulemaking. That rulemaking would focus on spot and cash digital commodity markets and address concerns about prediction markets and similar products.

Also Read: Asia has not opened yet: What will the first bell reveal about Bitcoin and oil?

Payment stablecoins would face a ban on passive interest or yield on idle balances. The bill would still allow activity-based rewards under negotiated language in Section 404. The White House and industry regard this compromise as settled. Bank lobbying continues. Developer protections remain. They aim to shield non-controlling software authors from being treated as money transmitters simply for publishing code.

If Congress enacts it, major DeFi and stablecoin projects operating in the United States would need clearer compliance strategies. Many pure-protocol developers could gain a safer legal footing. This mix could reshape compliance planning for large platforms while offering greater protection for developers who do not control user funds or protocol decisions.

The next step is a cloture vote on 15 September 2026. The motion requires 60 senators to agree to proceed. Republicans hold 53 seats. At least seven Democrats or independents must support it. Some Democratic aides call unresolved ethics disputes the biggest stumbling block. Law enforcement and banking groups push competing changes to ethics and stablecoin language. Prediction markets and research desks assign relatively low odds that the bill will become law in 2026. Failure at this stage could push comprehensive US crypto legislation back several years. That outcome would leave SEC and CFTC rulemaking as the main path for crypto regulation. Confidence in the text and vote timing is moderate because multiple independent reports agree. Political outcomes remain uncertain.

Also Read: Why a strong jobs report hit Bitcoin and Ethereum harder than the stock market

The revised CLARITY Act is the most concrete attempt so far to give US crypto markets a formal federal rulebook, especially for DeFi, exchanges, and stablecoins. Whether it advances on 15 September will shape how quickly US based projects can plan around stable rules rather than piecemeal regulation. Crypto users should watch that vote and any follow-up agency rulemaking as key signals for the regulatory regime they will face.

In my view, the current dip is a natural cooldown after a strong monthly rally, which a leverage unwind sharpened. The high correlation with traditional markets shows that crypto remains sensitive to macro liquidity sentiment. The near-term outlook is cautiously bearish. Momentum has shifted to sellers. Stability depends on holding the key support zone. Bitcoin must defend US$76,000 in the hours leading up to the CPI release. If it fails, macro fears could trigger another leg down. If it holds, the market may find room to consolidate and wait for clearer data.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.

Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.

The post Bitcoin drops to US$76,796.54 as 91% S&P 500 correlation exposes crypto’s macro trap appeared first on e27.

Southeast Asia’s venture boom has left behind a very large graveyard. Between January 2020 and July 2026, 7,538 startups across the region deadpooled, according to Tracxn data examined in a new e27 analysis of 76 notable venture-backed failures.

The pattern is unmistakable: 2021 and 2022 alone account for 57.3% of all closures, as founders who raised at peak 2019–2020 valuations ran out of runway once cheap capital vanished and investors began demanding proof of unit economics rather than growth alone.

E-commerce, social commerce, proptech, co-working, fintech, Web3 and logistics absorbed the heaviest losses — sectors where scale depended on constant cash injections rather than durable margins. Indonesian fashion platform Sorabel, Vietnamese proptech Propzy and enterprise AI firm Taiger all raised meaningful capital before folding. A second, smaller wave hit in 2024, when companies that had survived on bridge financing ran out of options as Series B and C capital failed to arrive.

Closures have since slowed sharply, to just 308 across 2025 and the first seven months of 2026, suggesting the most indiscriminate phase of the correction has passed. The lessons for founders: funding timing matters, GMV is not profit, and premature regional expansion punishes more than it rewards in a market that is really a patchwork of many.

Read the full article here:

—

REGIONAL

Grab in talks to acquire SoftBank-backed Atome: Grab is in discussions to buy buy-now-pay-later firm Atome, which counts SoftBank among its backers. A deal would significantly expand Grab’s fintech footprint across Southeast Asia’s fast-growing consumer credit market.

TikTok affiliate to sink US$980M into Vietnam logistics: The investment marks one of the largest single logistics commitments tied to e-commerce in Ho Chi Minh City, deepening TikTok’s Vietnam supply-chain footprint.

Circle to acquire Tazapay, expanding USDC’s SEA reach: The deal pulls Singapore-based cross-border payments infrastructure into Circle’s stablecoin ecosystem, signalling stablecoin rails moving deeper into SEA trade flows.

BillEase doubles credit facility with RCBC to US$16M: Philippine BNPL platform BillEase has doubled its credit facility with Rizal Commercial Banking Corporation to US$16M, boosting its lending capacity to serve more underbanked Filipino consumers.

Indonesia lawmakers grill Tokopedia, TikTok Shop over frozen funds: Parliamentarians are pressing the e-commerce platforms for answers on seller funds that remain frozen, raising fresh questions over merchant protections.

Malaysia ranks second to Singapore in GSMA digital index: The ranking cites Malaysia’s digital infrastructure and AI policy progress, with Vietnam, Thailand and the Philippines close behind in the region-wide assessment.

Philippine EV adoption set to accelerate despite charging gaps: BMI forecasts faster uptake even as charging infrastructure lags demand, pointing to policy and import incentives as key swing factors.

Singapore tops Southeast Asia in 2026 Best Workplaces list: The ranking places Singapore ahead of regional peerson workplace culture metrics, reinforcing its pull for regional talent.

Malaysia targets US$5B in annual GDP gains from AI: The communications minister linked the target to national AI strategy execution, part of a broader push to position Malaysia as a regional AI hub.

Singapore data centre investment to hit US$19.2B by 2050: PwC’s projection reflects sustained capacity build-out as demand for AI compute keeps climbing in the city-state.

N&E Innovations nets US$1.6M to fight cashew waste: The funding backs tech that converts husk waste into produce-protection material, tapping agritech’s growing climate angle.

Try Everything 2026 puts AI, robotics, capital centre stage: The event convenes cross-border investors and founders around emerging-tech themes shaping the region’s next funding cycle.

Toku expands into UAE, Talabat, eight Middle East markets: The SEA-founded fintech is pushing beyond its home region, following a broader pattern of SEA startups chasing Gulf expansion.

INTERVIEWS AND FEATURES

Over 7,500 SEA startups have shut down since 2020: New tracking shows a sustained capital drought has wiped out thousands of Southeast Asian startups over six years, underscoring how tight venture funding remains across the region.

SEA’s AI boom has a water problem it cannot offset away: The piece argues data centre cooling demands pose a resource strain that carbon-style offsets cannot easily solve for the region.

INTERNATIONAL

Massachusetts hits data centres with new clean power rules: Facilities over 25MW must now source their own clean energy or pay into a ratepayer fund, making Massachusetts the third US state in as many months to restrict data centre growth.

DOJ seeks more answers on Fox’s US$22B Roku deal: Regulators are pressing for further detail on the proposed acquisition, adding a layer of scrutiny to one of the year’s largest media-tech tie-ups.

Apple’s new CEO revives a Steve Jobs-era iPhone strategy: John Ternus used his first keynote as CEO to position the iPhone as Apple’s AI hub, echoing Jobs’ 2001 “digital hub” pitch for the Mac.

Apple unveils its first foldable, the iPhone Duo: The launch marks Apple’s entry into foldables, a category rivals have shipped for years, as the company leans harder into hardware differentiation.

Uber invests US$10M in Indian fleet operator Carrum: The Series B values Carrum at roughly US$168M post-money; the startup runs about 5,100 vehicles and is Uber’s largest fleet partner for its premium Black tier in India.

Google DeepMind alumni build tools to accelerate fusion power: The founders are applying AI-driven modelling to fusion reactor design, part of a wider wave of AI-for-energy startups chasing grid-scale power breakthroughs.

Travis Kalanick’s Atoms eyes a move into robotaxis: The Uber co-founder’s new venture is reportedly exploring autonomous ride-hailing, pitting Kalanick against the industry he helped build a decade earlier.

Students who avoid AI for schoolwork outscore peers: OECD: The finding holds true in science specifically, with one notable use-case exception, raising fresh questions for education policy as AI tools spread in classrooms.

Finvolve backs Indian cybersecurity startup AllSecureX: Singapore-based Finvolve has invested in AllSecureX, an Indian cybersecurity startup, signalling growing cross-border venture activity between Southeast Asia and South Asia’s expanding infosec sector.

SEMICONDUCTOR

OpenAI deepens Samsung tie-up on next-generation chips: The expanded partnership points to tighter supply-chain coordination between frontier AI labs and Korean chipmakers as compute demand accelerates.

Synopsys, A*STAR team up on AI chip packaging: The Singapore collaboration targets advanced packaging bottlenecks that increasingly constrain AI chip performance and supply.

China’s Longsys slips on Hong Kong debut: The memory chipmaker’s listing fell short of expectations, a reminder that not every chip-sector IPO is riding the AI hardware boom smoothly.

CYBERSECURITY

Hacker steals US$340M in crypto heist, returns most of it: The Liquid Network exploit ranks among the largest crypto thefts on record; roughly 85% of the stolen bitcoin has since been returned.

Google flags threat actors weaponising agentic AI: The warning covers attackers automating intrusion chains with autonomous AI agents, a growing concern for enterprise security teams across Asia.

AI

OpenAI adds AI doomer Paul Christiano to its board: Christiano, an alignment researcher who helped pioneer RLHF, joins OpenAI’s Safety and Security Committee following a string of AI agent security incidents.

Anthropic researcher quits, warns on self-improving AI: Jacob Coxon resigned publicly, accusing frontier labs of racing toward recursive self-improvement without adequate safeguards against loss of control.

OpenAI confirms wiki breach, pledges disclosure framework: The lab is working on transparency measures after a security incident, amid mounting scrutiny of how AI labs report agent misbehaviour.

Singapore firms race to adopt AI despite system gaps: A new report finds adoption is outpacing readiness, with many organisations lacking the underlying data and governance systems to support it.

THOUGHT LEADERSHIP

The cheapest way to stop your AI product from regressing: A contributor argues disciplined evaluation practicescost far less than the reputational damage of shipping AI features that silently degrade over time.

Quantum’s ChatGPT moment is coming, and it’s worth trillions: The piece contends quantum computing is nearing an inflection point comparable to generative AI’s breakout, with major economic stakes attached.

I built a 21-role AI workforce; management was the hard part: The founder’s account centres on orchestrating AI agents, arguing coordination — not capability — is the real bottleneck in agentic setups.

The missing layer in AI innovation: human verification: The essay makes the case that unchecked AI outputs need structured human review before they can be trusted in production systems.

When a slot opens, let the AI agent act within limits: The contributor argues for bounded autonomy — giving AI agents room to act inside clearly defined guardrails rather than full or no autonomy.

The most valuable part of AI may not be the model: The essay suggests the surrounding infrastructure — data, workflows, and integration — may capture more long-term value than the models themselves.

Gen Z doesn’t need more AI courses, it needs unreplicable skills: The argument centres on judgement and creativityas the skills that will differentiate young professionals once AI commoditises technical execution.

SEA’s SMEs don’t have an AI problem, they have a record problem: The piece argues poor data hygiene, not AI access, is the real barrier holding small businesses back from useful automation.

Singapore learnt to measure cooling; ASEAN should follow: The contributor argues Singapore’s cooling metrics offer a template the rest of ASEAN should adopt as data centre density rises regionally.

Thailand’s mobility future will be decided by data, not vehicles: The essay argues data infrastructure, not fleet electrification alone, will determine how Thailand’s transport systems evolve.

Will Bitcoin hold US$77,000, or slide toward US$2.51T market cap?: The analysis lays out scenarios for bitcoin’s trajectory heading into 10 September, weighing support levels against broader market pressure.

Asia hasn’t opened yet: what the first bell may reveal: The piece previews how Asian markets might react to overnight moves in bitcoin and oil prices as trading begins.

Future-proofing the past: smart engineering for ageing condos: The essay explores how retrofitted infrastructurecan bring Singapore’s older condominiums up to speed for EV charging demand.

The post Ecosystem Roundup: What failed startups reveal about SEA’s next tech cycle appeared first on e27.