Olive AI raised US$902 million, deployed automation across 900 hospitals in 40 states, and was valued at US$4 billion at its 2021 peak. By October 2023, it was gone. Not because the AI failed — but because a routine Epic module update broke the bots, and hospitals found themselves adding human monitoring on top of the automation they’d paid to replace. A minor interface change. A catastrophic systems mismatch. The product worked perfectly in the lab. The lab was not where nurses actually worked.

Pear Therapeutics received FDA approval for prescription digital therapeutics, then collapsed because doctors had no workflow to prescribe them, pharmacies had no system to fulfil them, and insurers had no billing code to reimburse them. The product existed in a system vacuum. Babylon Health scaled AI diagnostics to millions of users, then watched clinicians run parallel manual checks on every AI output — functionally doubling the workload it was built to eliminate. Both companies shut down in 2023.

These weren’t technical failures. They were a category error: treating a workflow intervention as a product launch.

Bain’s 2024 analysis of over 900 companies found that 88 per cent of business transformations fail to achieve their original ambitions. IDC puts the annual cost at US$2.3 trillion, more than the GDP of Italy, gone every year from systems that worked and were still rejected. The common finding across McKinsey, BCG, and Harvard Business Review is consistent: the failure driver is not a technical limitation. It is the gap between what a product can do and how people actually work. This is not a medtech story. It is not an agritech story. It is the story of every industry where a human being stands between your technology and its purpose, which is to say, every industry that exists.

The real constraint is human bandwidth, not computation

Ask a clinician why they rejected a diagnostic tool with 15 per cent better accuracy. The answer is almost never distrust of the algorithm. It is four minutes. Four extra minutes per patient is trivial in isolation. Across a 12-hour shift with 30 patients, it is catastrophic — especially when the pharmacy call, the handover note, and the attending physician’s interruption are all queued behind it.

The same dynamic plays out in a fulfilment warehouse where a new routing system adds two extra taps per package scan. In a law firm where a contract review tool requires a different login than the document management system. In a retail bank where a fraud-detection upgrade changes the screen flow that branch staff have navigated by muscle memory for six years. The technology improves the outcome. The friction destroys the adoption.

Also Read: Bridging the AI trust gap: Why ad diversification and creative differentiation are the future of customer connections

Research published in NEJM Catalyst identifies the primary barrier to clinical technology adoption not as accuracy distrust but as muscle memory disruption. The same principle holds everywhere humans operate under time pressure and cognitive load, which is most places where technology is now being deployed. Mayo Clinic created a new executive role — Chief Clinical Systems and Informatics Officer — specifically because hospitals now push hundreds of software changes per quarter to clinical staff. Each one is a tax on attention. Accumulate enough taxes, and the immune system activates: staff quietly revert to what they know, regardless of what the trial data showed.

What IDEO actually did, and why it matters

When the American Red Cross hired IDEO to address declining blood donation rates, the instinctive solution would have been to optimise the process: faster check-in, better needles, shorter waits. IDEO did something different. They observed.

What they found was not a logistics problem. It was an emotional one. Donors came in anxious about the needle and left feeling like a transaction. The post-donation routine — sit for 15 minutes, drink juice, eat a cookie — treated recovery as a waiting room problem. Nobody asked what had brought the donor in. Nobody made the moment mean anything.

IDEO’s intervention was not a product. It was a ritual redesign. During the post-donation observation period — when donors had to remain seated anyway — staff handed them a card and a pen and asked them to write down why they had donated. Not for a form. Not for a database. Just to hold, and to keep. The cards were photographed and pinned to a display board in the donation centre, as Post-it notes from a first date, casual and personal and visible to the next person who walked in.

The intervention cost almost nothing. It changed the emotional architecture of the entire experience. Donors who had articulated their own motivation — in their own handwriting, in their own words — returned at dramatically higher rates. They hadn’t just given blood. They’d made a statement about who they were. The Red Cross hadn’t improved the needle. They had changed what the act meant.

The breakthrough wasn’t a better product. It was the recognition that behaviour follows meaning — and meaning, if you design for it, can change everything

This is what anthropological design actually is. Not user research as a checkbox before engineering starts. Not a UX audit after launch. It is treating technology, behaviour, and context as a single system — where the human workflow is not a constraint to be managed around, but the primary design surface.

Over three decades as an investor, I’ve listened to thousands of founders and CEOs explain their technology, their product roadmaps, and their market strategies. Almost none could articulate how customer experience would be architected over time — or how workflow integration would be continuously tested, retrained, and adapted as real-world conditions evolved. The implicit assumption was always the same: build the product, and adoption will follow. It rarely does. And the gap between that assumption and reality is where most of the US$2.3 trillion goes.

Also Read: Solving multiple medtech problems with a single device powered by AI

The companies getting this right

John Deere’s See & Spray — built on a US$305 million acquisition of computer vision startup Blue River — uses AI to identify weeds and spray only them, cutting herbicide use by up to 77 per cent. It could have been another brilliant system ignored in a barn.

Instead, Deere built the adoption architecture before the product reached the market: three pricing tiers structured around farmers’ capital constraints, software designed so existing precision-ag users were “more than halfway to full autonomy” before touching a new feature. Deere’s CFO framed the goal explicitly as meeting farmers “at every stage of their precision tech journey.” By the end of 2024: record adoption across the entire technology stack.

The insight is not complicated. Farmers — like clinicians, like warehouse workers, like anyone operating under time pressure in a high-stakes environment — are not resistant to technology. They are resistant to discontinuity. Products that require behavioural rupture fail. Products that slip into existing rhythms without announcing themselves compound quietly until they’re indispensable.

Organisations with a structured change-management strategy are seven times more likely to meet their digital transformation goals, per Mendix’s analysis. Not from a better model. From understanding the human system, the model is entering.

A question worth sitting with

The US$2.3 trillion graveyard of failed transformations is not primarily an engineering failure. It is a failure of scope — a discipline that stopped at the boundary of the product and called it done. Superior technology is necessary. It is no longer sufficient. The bottleneck has migrated from the lab to the deployment environment, and most organisations are still staffed for the old bottleneck.

So here is the uncomfortable question — not just for medtech founders or agritech operators, but for anyone building anything that a human being will eventually have to use, adopt, or trust:

If you cut your technical team in half tomorrow and replaced them with anthropologists, ethnographers, and workflow specialists, would your product get worse?

If the answer is no — or if you’re not sure — that uncertainty is the finding. The next defensible moat will not be built in a model. It will be built on the accumulated institutional knowledge of how adoption actually works — knowledge that compounds with every deployment, cannot be licensed, and cannot be replicated from a term sheet.

Start there.

This article is part of David Kim’s Value Creation column. It sits alongside the Asia Value Creation Awards, which aim to recognise PE and VC teams driving long-term, fundamentals-led value creation across the region.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. Share your opinion by submitting an article, video, podcast, or infographic.

Enjoyed this read? Don’t miss out on the next insight. Join our WhatsApp channel for real-time drops.

The post Value creation: Your US$900M AI is failing because humans don’t work the way you think appeared first on e27.

Most Board packs on cyber, privacy, vendor exposure, and resilience fail in the same way. They contain activity, metrics, and updates, yet still leave senior decision makers unclear on what they are being asked to govern.

One page shows phishing rates. Another shows patching. Another shows third-party incidents, privacy breaches, or resilience testing. Each page may be accurate on its own, but the Board still leaves without clear answers to the questions that matter. What could disrupt the bank’s most important services? Where customer harm is most likely to emerge. Which dependencies have become strategically dangerous? Which weaknesses can be tolerated for now, and which require action before the next incident forces the decision?

Boards rarely suffer from too little information. They suffer from information organised around functions rather than decisions.

Supervisory expectations increasingly point in the same direction. Boards are expected to understand important business services, consider severe but plausible disruption, receive timely reporting on material weaknesses, and use that reporting to make investment and risk decisions. That is not a standard built for fragmented dashboards. It is a governance standard built for judgment.

The mistake is to report by domain instead of by consequence

Most institutions still report cyber, privacy, vendor, and resilience as separate disciplines.

The Board does not govern those areas as isolated territories. It governs the bank’s ability to operate safely, protect customers, withstand disruption, and remain within risk appetite. Once reporting is divided into specialist slices, the most important relationships disappear. A third-party weakness no longer looks like a resilience issue. A privacy control gap no longer appears connected to cyber exposure. A resilience weakness no longer looks like a conduct issue. The Board receives a set of departmental truths rather than one decision grade view of institutional risk.

Also Read: ESG as strategic value: Why Asian boards must move beyond disclosure

A Board paper should answer one question

What are we being asked to decide?

A Board does not need another description of open high-severity vulnerabilities unless that information is linked to a consequence it can govern. It does not need a recital of privacy incidents unless management can explain whether those incidents point to weak design, poor third-party control, weak customer communication, or a deeper failure in data stewardship. The same applies to resilience testing. The governance question is not simply whether a test happened, but whether the outcome changes management’s confidence in staying within impact tolerances for important business services.

The strongest Board narratives, therefore, start with business consequence, not control category. They begin by showing which services, customer outcomes, regulatory obligations, or strategic dependencies are at risk. Only then do they explain which cyber, privacy, third-party, or resilience factors are driving that exposure. The order matters because it forces management to translate control data into a decision about risk acceptance, investment, sequencing, or intervention.

What a joined-up Board narrative should contain

First, it identifies the service or outcome that matters. Not a generic technology issue, but a business service, customer process, regulatory duty, or strategic dependency that the Board would recognise as material.

Second, it shows the chain of exposure. This is where cyber, privacy, vendor, and resilience become one story. A critical service may depend on a concentrated third party, a weak privileged access model, poor data lineage, or an untested recovery path. A privacy issue may be the downstream result of weak identity governance, excessive access, or poor vendor oversight. The Board needs to see the chain, not just the symptom.

Third, it sets out management judgment. What is already being done? What is improving? What remains outside the target state? What assumptions is management making? Where confidence is high and where it is not.

Fourth, it states the decision required. Does the Board need to support a risk acceptance, a control uplift, a delay to a strategic initiative, a change in tolerance, or a sharper intervention on execution? Without this final step, the pack informs but does not govern.

A credible challenge depends on narrative quality

Boards are often told they must provide effective challenge. That is true, but incomplete. A Board cannot challenge credibly if management presents risk through a structure that obscures cause, consequence, and uncertainty.

Directors then end up asking weaker questions. Why is the number red this month? Why is this metric worse than last quarter? Why has this vendor issue not been closed? Those are reasonable questions, but they do not reach the real issue when several risks are combining to threaten a major service or customer outcome.

Also Read: The always-on boardroom: When strategy stops being an event

This is why Boards need fewer comfort metrics and more explicit statements of uncertainty. Where is management relying on vendor attestation rather than direct evidence? Which recovery assumptions have not been tested end-to-end? Which privacy controls look compliant on paper but remain weak in practice? Which cyber improvements reflect genuine resilience, and which simply reflect better measurement? These are the questions that improve governance.

Third-party and privacy reporting need business language

One of the biggest weaknesses in Board reporting is the way third-party risk is still presented as a procurement topic when it is increasingly a strategic resilience issue. A Board does not need a longer supplier inventory. It needs to understand where concentration, substitutability, recovery dependency, and service integration create fragility in the bank’s ability to deliver important services.

The same logic applies to privacy. Privacy reporting often becomes either legalistic or reduced to incident counting. Both approaches are too weak. A stronger approach is to report privacy as a question of trust, customer treatment, and decision quality. Are we using customer data in ways we can genuinely defend? Are controls reducing operational data sprawl or merely documenting it? Are cyber weaknesses, poor access design, or third-party handling creating conditions for privacy harm at scale?

What Board ready reporting should feel like

A good Board paper should leave directors able to answer a small number of hard questions with confidence. Which important services and customer outcomes are under the greatest pressure? Which dependencies and control weaknesses are creating that pressure? Which issues management is handling, and which require Board support or intervention. Where is the institution relying on an assumption rather than proof?

Also Read: What to actually prioritise when your board wants AI and everything feels urgent

Report in the language of consequence. Show the chain from cause to business impact. Make uncertainty visible. Connect control issues across domains. End with the decision management is really asking the Board to make.

If a pack cannot do those things, it is probably not Board-ready, no matter how polished the metrics may look.

Final thought

The future of governance in banking will not be won by institutions that collect the most cyber, privacy, vendor, and resilience data. It will be won by institutions that translate those issues into clear choices about service continuity, customer trust, risk appetite, investment, and management accountability.

Boards do not need another pile of fragmented indicators. They need a coherent narrative that tells them what matters, why it matters now, how confident management really is, and what decision is needed before the next disruption turns an unmade choice into a visible failure.

That is what decision-grade governance looks like.

—

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 build a board paper that actually answers: ‘What are we being asked to decide?’ appeared first on e27.

Global financial markets are currently undergoing a severe recalibration as the artificial intelligence trade unwinds. This paradigm shift is triggering a broad rotation out of high-flying momentum stocks and into defensive sectors. The contagion is evident across major Western indices. The S&P 500 retreated by 1.4 per cent to settle near 7,375, while the technology-focused Nasdaq Composite suffered a sharper 2.2 per cent decline. The Dow Jones Industrial Average demonstrated relative resilience, slipping a mere 0.09 per cent. Across the Atlantic, European markets also felt the pressure, with Stoxx 600 futures dropping approximately 0.9 per cent as they pulled back from recent record peaks.

The correction hit the Asia-Pacific region with exceptional force, driven by a sharp rout heavily weighing down high-flying technology and semiconductor firms. The MSCI regional benchmark plummeted 2.9 per cent. South Korea experienced the most dramatic fallout, with the KOSPI plunging roughly 10 per cent and triggering an automatic 20-minute trading halt. This massive wipeout was spearheaded by memory chip giants SK Hynix and Samsung Electronics, both of which cratered by over 12 per cent.

Japan saw the Nikkei 225 fall 3.6 per cent to close at 69,788.38, breaking below the critical psychological threshold of 70,000. In Greater China, the Hang Seng Index dropped 1.8 per cent to 23,445, cementing a bearish head-and-shoulders technical pattern, while local artificial intelligence software names like MiniMax tumbled 16 per cent intraday. The mainland saw the Shanghai Composite ease 1.4 per cent to 4,106 points, and the technology-reliant Shenzhen Component shed 3.2 per cent.

Beyond equities, the risk aversion sentiment extended to commodities and private technology valuations. Global oil prices retreated as geopolitical tensions in the Strait of Hormuz cooled, sending Brent crude down over one per cent to near US$76.95. In the technology sector specifically, Alphabet dived five per cent, and private aerospace titan SpaceX experienced a massive 16 per cent valuation crash. Investors are aggressively booking profits and pivoting out of growth areas into defensive pockets of the market, including select European semiconductor plays and financial institutions.

Also Read: From frontier to emerging: How Vietnam’s stock market rewrote the ASEAN playbook in 2025

This massive unwinding of the technology trade has created a direct spillover effect into digital assets, proving once again the tight correlation between traditional technology markets and cryptocurrency. Bitcoin has lost its clear upward direction and is currently wobbling in the US$62,000 to US$62,500 range.

The cryptocurrency broke key support levels two times during the Asian session before attempting to consolidate near US$62,370. Crypto buying power remains heavily constrained by stalled United States exchange-traded fund inflows and broader market anxieties regarding upcoming Federal Reserve monetary guidance.

The underlying catalyst for this synchronised selloff is a fundamental reevaluation of Federal Reserve interest rate expectations, accompanied by a slight spike in United States Treasury yields. Investors are aggressively pricing in the potential for a rate hike, forcing a rapid rotation out of growth assets. Market sentiment has turned decidedly bearish in the short term. This shift has triggered active prediction hedging on platforms like Kalshi and Polymarket, where speculative volume is surging as traders place bets on whether Bitcoin will test lower handles around the US$58,000 mark.

At the point of writing, Asia market has not started. Let’s see if it will go down further.

—

The post From silicon to satoshis: Tracing the contagion of the global market unwind appeared first on e27.

Vietnam is preparing to test a more interventionist model for building technology companies, with the Ministry of Science and Technology proposing a National Venture Capital Fund with initial capital of US$100 million for the 2026-2028 period.

The proposal, discussed at a meeting of the government’s science, technology, and innovation committee on Tuesday, is designed to accelerate the commercialisation of strategic technologies and support the creation of competitive technology enterprises.

Also Read: Vietnam isn’t just inviting private capital in. It is structurally dependent on it

While US$100 million is modest by global venture capital standards, the plan is significant for Vietnam and the wider Southeast Asian startup ecosystem. It signals Hanoi’s intent to move beyond policy support and infrastructure building and towards direct participation in venture financing, particularly in sectors where private investors remain cautious because of long development cycles, uncertain exits, and high technical risk.

The fund is being modelled on Israel’s Yozma programme, one of the most cited examples of state-backed venture capital catalysing a private VC industry. Under the Vietnamese proposal, the state would lead the fund’s initial capital structure. From 2028 to 2035, the fund would gradually mobilise more private capital, with private investors expected to account for 30 to 40 per cent of total capital. The ministry has also proposed the development of separate funds for different technology sectors.

Vietnam’s deeptech funding gap

The timing is notable. Vietnam has emerged as one of Southeast Asia’s more dynamic startup markets, supported by a young digital population, a growing engineering workforce, and increasing interest from regional and global investors. However, much of the country’s startup funding has flowed into consumer internet, fintech, e-commerce enablement, and software businesses rather than harder technology categories.

That is not unique to Vietnam. Across Southeast Asia, deeptech, advanced manufacturing, semiconductors, climate technologies, biotech, and other research-heavy sectors often struggle to secure early-stage risk capital. These companies typically require longer gestation periods, specialised evaluation, patient funding, and stronger links between universities, laboratories, corporates, and investors.

Vietnam is trying to position itself more aggressively in strategic technologies as global supply chains shift and as multinational technology companies expand their presence in the country. The country has already attracted attention as a manufacturing base for electronics and is attempting to move up the value chain into higher-value technology development.

A national venture capital fund could help bridge the gap between research and commercialisation, especially if it can back companies emerging from universities, research institutes, and incubators. But the challenge will be turning a state-funded vehicle into a credible venture investor rather than another public-sector grant mechanism.

The Yozma inspiration and its limits

Israel’s Yozma programme, launched in the 1990s, helped seed the country’s venture capital industry by using government capital to attract foreign and domestic private investors. Its structure gave private investors strong incentives and helped create a commercially disciplined investment culture.

Also Read: Vietnam’s biggest PE bet of 2025 was not on tech. It was on what 100M people eat every day

Vietnam’s proposal borrows from that logic: state capital comes first, private capital follows, and specialised funds are created around priority sectors. In theory, this allows the government to absorb some early risk while encouraging private investors to participate once the model matures.

But transplanting Yozma-style models is rarely straightforward. Israel already had strong research universities, defence-linked technology capabilities, global diaspora networks, and deep connections to US capital markets. Vietnam has different institutional realities, including a younger VC ecosystem, fewer proven deeptech exits, and a capital market still developing mechanisms for valuing high-growth technology businesses.

The ministry appears aware of these constraints. It has identified the tension between venture capital’s risk-tolerant nature and the public-sector principle of preserving state capital as a major challenge. That tension is central to whether the fund can function effectively.

Risk cannot be managed deal by deal

Venture capital works because a small number of outsized winners compensate for many failures. Public capital management, by contrast, often penalises losses on individual investments, even when the overall portfolio performs well. If officials managing the fund are exposed to personal or legal liability for failed startup investments, the vehicle could become too conservative to achieve its purpose.

To address this, the ministry has proposed that fund performance be evaluated across the entire portfolio rather than on individual deals. It has also called for protection mechanisms for decision-makers who follow proper procedures.

This is a crucial point. Without such protections, fund managers may avoid genuinely risky strategic technologies and instead back safer, later-stage, or politically favoured companies. That would undermine the rationale for creating a venture fund in the first place.

The ministry has recommended that the government report to the National Assembly to issue a resolution creating a specific mechanism for the fund. This would include liability exemptions for officials managing state-funded venture capital, provided they comply with regulations. The goal is to enable controlled risk-taking in investments involving strategic technologies.

Governance will decide credibility

The proposed governance structure also points to lessons from past state-backed investment efforts in the region. The ministry has recommended market-based recruitment and compensation, autonomy for the fund’s investment council, and stronger ties with research institutes, universities, and technology incubators.

These details are important. A venture fund needs experienced investors, sector specialists, and the ability to make decisions quickly. If compensation is not competitive, the fund may struggle to attract talent from the private market. If investment decisions are too bureaucratic, promising startups may look elsewhere for capital.

Autonomy will also be closely watched by private investors. For the fund to crowd in capital rather than crowd it out, it must be seen as commercially disciplined, transparent, and free from excessive administrative intervention.

Southeast Asia has no shortage of government-backed funding initiatives, from Singapore’s deep pool of state-linked capital to Malaysia’s startup financing schemes and Indonesia’s efforts to mobilise domestic capital for technology and innovation. The strongest models tend to combine public-sector strategic direction with professional investment management and clear accountability frameworks.

Also Read: Why Vietnam is the next big thing for startups and corporate partnerships

Vietnam now appears to be moving in that direction, but execution will be decisive.

The exit problem

Perhaps the most difficult issue is not capital deployment but capital recovery. The ministry highlighted Vietnam’s underdeveloped exit ecosystem, including limited mechanisms for valuing technology companies and insufficient channels for investors to recover capital.

This is a broader Southeast Asian problem. IPO markets remain uneven for technology companies, M&A activity is still limited compared with the US or China, and many regional startups depend on later-stage funding rounds rather than clear exit pathways. For deeptech companies, the problem is even more acute because buyers are specialised and commercialisation timelines can be long.

If Vietnam wants private investors to account for 30 to 40 per cent of the fund’s capital in later phases, it will need to improve exit visibility. That could involve strengthening domestic capital markets, encouraging corporate acquisitions, creating clearer valuation standards, and deepening cross-border links with regional and global investors.

The proposed US$100 million fund is therefore not just a financing instrument. It is a test of whether Vietnam can build the institutional architecture required for a more sophisticated innovation economy.

If designed well, the fund could help turn public research into commercially viable companies and give Vietnam a stronger position in Southeast Asia’s emerging deeptech landscape. If designed poorly, it risks becoming another state capital vehicle constrained by caution, weak incentives, and limited exits.

For now, Hanoi has identified the right problems. The harder task will be building a fund that is allowed to take the risks venture capital requires.

The post Vietnam looks to Israel’s Yozma model for US$100M national venture fund appeared first on e27.

As AI blurs the boundary between human and machine online, a new platform is making the case that proving you are a real person should be as fundamental and private as having a password.

World ID, developed by Tools for Humanity, is a digital identity credential designed to verify that a user is a unique human being without collecting or retaining personal information. The platform is now expanding across Asia, with Singapore serving as a key regional hub.

At its core, World ID functions as a modern proof of personhood. Andrew Hsu, General Manager for Singapore and Taiwan at Tools for Humanity, describes it in an email interview with e27 as “a modern-day blue checkmark for personhood” — one grounded not in celebrity or social standing, but in the simple fact of being human.

The practical applications span a wide range of industries. Concert ticketing platforms can use World ID to ensure tickets reach genuine fans rather than automated bots. Dating apps can confirm that profiles belong to real people. Enterprise tools can verify that the individual on a video call or behind a digital signature is who they claim to be. Partners already integrating the tech include Tinder in Japan, Zoom, DocuSign and Concert Kit, which has announced plans with the band 30 Seconds to Mars.

The bot problem World ID is built to solve

The platform emerges at a moment when traditional verification methods are under significant strain. Bots can now bypass CAPTCHAs more reliably than humans, and AI-generated documents are increasingly capable of deceiving legacy systems. Fraudulent accounts, deepfakes, and synthetic identities are no longer edge cases. They are, according to Tools for Humanity, a growing structural threat to digital trust.

Also Read: Vietnam looks to Israel’s Yozma model for US$100M national venture fund

“As AI advances, it is blurring the line between human and machine interactions online,” Hsu says, “making it harder to prove with certainty that a person is truly a person.”

World ID addresses this by separating the question of identity — who you are — from the question of personhood — that you are human.

How the Orb works

Verification is conducted through a device called the Orb. When a user downloads the World App and presents themselves at an Orb station, the device photographs their face and irises. These images are used to generate an encrypted code, which is split into randomised fragments and sent to the user’s device before being permanently deleted from the Orb.

The encrypted fragments are then compared across independent compute nodes run by third parties including universities using a process called Anonymised Multi-Party Computation. This confirms that the individual has not previously verified without revealing who they are. Neither World nor Tools for Humanity retains any personal information from the process, and users may delete their data at any time.

Independent security audits of the Orb and its software have been conducted by cybersecurity firms Trail of Bits and Theori, with results made publicly available. The underlying tech is open-source.

Also Read: Who am I in the age of AI? Identity, displacement, and awakening

Singapore as a regional blueprint

In Singapore, Orb stations have been deployed at self-serve locations through partners including Collin’s and Sakae Sushi restaurants, with community pop-ups also under way. Hsu describes Singapore as a “bellwether” for Southeast Asia, citing its strong public-private collaboration and high awareness of digital safety issues.

The company acknowledges that building public confidence will take time. Hsu frames education, accessibility and regulatory engagement as the three pillars of its approach to new markets — noting that trust, ultimately, “is earned over time and through consistent action.”

World ID is currently available to users aged 18 and above.

—

Image Credit: World ID

The post Can World ID solve the internet’s fake human problem? appeared first on e27.

When Meta announced that CRED founder and one of India’s most celebrated fintech entrepreneurs, Kunal Shah, would become the global CEO of WhatsApp, the Indian internet went predictably delirious. LinkedIn filled with tributes. Venture capitalists took victory laps. Across Southeast Asia, the story landed as another chapter in the South Asian diaspora’s march through Silicon Valley’s upper echelons.

But after more than a decade covering startup ecosystems across India and Southeast Asia, I have learnt one thing: the most important story is rarely the one generating the most LinkedIn posts.

The acquisition behind the appointment

The sequencing here is enormously crucial, and most celebratory coverage has glossed over it entirely. Meta did not simply hire Shah but it acquired his loyalty and credit-card management platform CRED for US$4.5 billion, with US$900 million injected as fresh capital. The CEO title came bundled with the deal. These two events are inseparable.

Also Read: China blocks Meta’s AI bet on Manus: What it means next

Founded in 2018, CRED raised over US$1 billion in VC funding and posted net losses of approximately US$175 million in its most recently reported financials. Its core product, a rewards platform for credit card holders, was never a conventional revenue-generating business. What it built, meticulously over the years, was something far more valuable to Meta: roughly 25 million verified, curated profiles of affluent, creditworthy Indians.

At US$4.5 billion, that works out to approximately US$180 per user or, stripped to essentials, around US$36 per high-quality financial profile. Meta did not buy an app but bought a dataset and a monetisation shortcut.

WhatsApp’s long-standing India problem

To understand why that dataset matters so urgently, consider WhatsApp’s peculiar India paradox. The county is WhatsApp’s largest market, accounting for approximately 26 per cent of its global user base, with around 15 million active WhatsApp Business accounts. By every measure of adoption, India is a triumph.

By when it comes to revenue, it is an embarrassment. WhatsApp contributes less than 2 per cent of Meta’s total global revenue, and India’s contribution to even that meagre figure is disproportionately small.

WhatsApp Payments, launched in India in 2018 amid breathless predictions that it would render the entire Indian fintech industry obsolete, never came close to delivering. PhonePe and Google Pay dominate UPI transactions. The failure was never about Indian consumers — they adopted digital payments with extraordinary enthusiasm — but about Meta’s inability to commit to the local execution focus the market demanded.

Also Read: Meta × Manus: The misread AI deal

That is the mandate Shah has actually been handed: fix the India monetisation problem, then export the playbook to Indonesia, Brazil, Nigeria, and every other large emerging market where WhatsApp dominates daily communication.

The question nobody is asking

This is where the celebration deserves serious scrutiny.

If CRED’s core asset is 25 million verified financial profiles of affluent Indians, what exactly happened to those profiles when the acquisition closed? India’s Digital Personal Data Protection Act 2023 requires explicit user consent before personal financial data is transferred to any third party. It restricts cross-border data transfers to countries on an approved whitelist, a list that remains unfinished, with the United States not yet on it.

Did CRED’s 25 million users individually and knowingly consent to their credit profiles being transferred to an American technology conglomerate? Or did nobody check?

Not a single Indian regulator publicly raised this question, nor a parliamentary inquiry was filed. The mainstream Indian press was too busy writing about a middle-class founder’s inspiring journey.

The contrast with how China handled a structurally similar situation is striking, not because Beijing’s authoritarian methods are worthy of admiration, but because the underlying principle is worth acknowledging.

When Meta reportedly invested US$2 billion in Manus AI, a Chinese startup that had relocated to Singapore, Beijing forcibly unwound the deal and called it a national security matter. Citizen data, it declared, is sovereign infrastructure. One need not endorse Beijing’s governance to recognise that this position is increasingly mainstream in serious technology policy circles, from Brussels to Singapore itself.

What the ecosystem is choosing not to see

Shah’s earlier venture, Freecharge, was sold to Snapdeal in 2015 for approximately US$400 million. Snapdeal offloaded it to Axis Bank two years later for around US$53 million, an 87 per cent write-down. As recently as FY25, Freecharge was still posting net losses of approximately US$5.6 million. One LinkedIn commentator put it bluntly: the business was never real. The exit was.

CRED followed a similar arc. Enormous capital raised, persistent losses, and then a multi-billion-dollar exit priced not on business fundamentals but on the future value Meta believes it can extract from the underlying data.

Shah, to his credit, is one of Indian startup culture’s more intellectually honest voices, and his elevation carries genuine symbolic weight. The critique here is not personal but systemic. He played by the rules the ecosystem created. The real question is why nobody changed those rules.

Two cheers, with conditions

For readers across Southeast Asia, the lesson is clear: the most consequential technology policy is the kind enforced before a deal closes, not debated after the data has already moved.

Also Read: Autonomous agents in performance marketing: A critical look at Meta’s US$2B Manus AI

Shah may yet prove to be exactly the leader who turns WhatsApp into the financial services giant Meta desperately needs. That possibility is genuinely exciting. But enthusiasm is not a substitute for accountability. And a good story is not the same thing as the whole story.

The post WhatsApp’s new CEO is the headline. India’s data is the story appeared first on e27.

What does it take to tell a winning story? Through my experience working in public relations, insights management, and tech entrepreneurship, I’ve come to see that storytelling isn’t a soft skill. In 2026 and beyond, it’s analytical output.

The big companies know this. Job listings mentioning “storytelling” as a skill or headlined with titles like Head of Storytelling or Narrative Strategist now feature on Indeed at a growing rate, instead of the traditional communications manager or head of marketing titles. Job postings featuring the word “storyteller” have grown significantly.

Rapid growth of AI content has shifted messaging into a new territory: one where good-quality messaging has a chance to stand out in a pile of ever-growing AI slop.

Through my own successes and failures in business and entrepreneurship, I’ve realised that to communicate persuasively, to investors, customers, or partners, you cannot rely on surface-level messaging. When many businesses compete in trying to tell the same story, sell similar products, and stand out in a cluttered market, my advice is to follow one simple rule: be the most reliable.

For SMEs and startups, this can seem challenging. How can a small business seem more reliable than a large player?

How do we stand out in the noise of everyone yelling, “I’m the best!”?

The rise of AI sloppytelling

The harsh truth is that most businesses were not strong communicators to begin with, so many of us now turn to AI to write a convincing pitch deck, transform our slide headlines into winning mantras, or write our business plans for us: the one-click-win. We’ve come to depend on it because AI messaging is better than no messaging, right?

As a result, inauthentic AI-devised content is taking over. Much of this output lacks focus and conviction to make audiences feel at ease. A generic story doesn’t do the job of putting prospective customers at ease, especially when your competitor is larger, has bigger teams, a longer runway, or more experience.

Using AI to write your story will not make you reliable or convincing. It will make you unremarkable. I call it Sloppytelling!

AI produces stories that often lack impactful key messages. The one-liners that make people stop and think, evoke nodding heads, and build confidence in your offer. This is because AI is based on analysis of past collections of vast generic data sets, and your story is based on your data and yours alone. It comes from the conversations you’ve had with clients who signed on the dotted line, and from those who opted not to.

Also Read: The storytelling myth: Why narrative-first leadership is overrated

AI is also not a replacement for real stakeholder or customer feedback. The actual data about your product or service shapes the story you tell. Remember: storytelling is analytical.

Finally, I argue that you should not use AI for developing well-structured responses to tough questions. Tough questions almost deserve their own section here, but we know them well from job interviews, investor meetings, and our business development presentations. It’s the questions that broke us.

Strong storytelling is as much about creating those powerful one-liners as it is about building our defensive comebacks, the fortress of words that protects our business from scrutiny.

A framework for analytical storytelling

Storytelling that can change minds or close deals is usually built on lessons accumulated over time, often through difficult experiences with customers and stakeholders. We start to form an idea of the optimal story through our repeated encounters with the word “no.”

After nearly two decades in business, listening and learning from strong storytellers, my consultancy focuses, among other things, on helping SMEs learn to tell their stories well and build lasting connections through their words and messaging. I argue that there are ways to bypass the painful rejections, the time spent hearing our least favourite word, through approaching storytelling in a structured and analytical way.

As mentioned earlier, the first part of this is to focus on insights, specifically around answering three questions about your business:

• What makes our proposition unique?
• Why do we do this better than others?
• Which questions will kill our business?

While most companies focus on the first two questions, many fail to focus on the final question. This one is complicated because there are many questions that can kill us, and our stakeholders ask different questions based on their needs or concerns. You have strong storytelling when your pitch incorporates the answers to these questions, not just once, but repeatedly, until this is what your audience remembers.

How AI can support your storytelling

You should certainly use AI to build your story! The tools are there to make our work more efficient and save us time. The strongest value AI can provide is to support the analytical work and insights generation that feeds your story, the legwork you need to do before formulating your winning pitch. It can also help dot the i’s and cross the t’s after your story is written.

Also Read: How brands are crafting communities through the art of visual storytelling

Here is how I propose you use AI in your storytelling process:

• To get clarity into your business and support analytical insights
• To filter your content and spot patterns behind your key messaging
• To devise questionnaires for stakeholders or customers to uncover weaknesses
• To clean up grammar and phrasing

AI can save you time and handle some of the legwork, but it can’t get you all the way there in answering the three core questions.

The takeaway

Many people still think of storytelling as something soft, artistic, or nostalgic: a campfire, a childhood book, a good writer’s craft. But in business, storytelling is not a cosy blanket. It is the bare-bones framework that holds your relationships with stakeholders, customers, and prospective clients in place. For SMEs and startups, it can be one of the most powerful tools for competing with larger players.

In a market saturated with generic, AI-assisted messaging, the winners will not be the companies producing the most content. They will be the ones whose stories make a lasting impact. That starts with investing in analytical storytelling today.

—

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 Why storytelling is no longer a soft skill but analytical output appeared first on e27.

Everyone wants to build the pipes. OpenAI, Stripe, Google, Visa — the race to let AI agents spend money autonomously is well funded and well covered. What is conspicuously absent is the layer that asks a far more uncomfortable question: should the agent be spending that money at all, and can the data driving that decision actually be trusted?

Telegraph Protocol’s Mark Basa and Ahmed Ali are not wrong to raise this. Hallucinations in a chatbot are a UX annoyance. Hallucinations in a system authorised to execute financial transactions are a liability event. And when those errors happen at machine speed, across multiple merchants simultaneously, the damage compounds before any human can intervene.

The liability question is equally unresolved. Existing legal frameworks assume human intent somewhere in the chain. Autonomous agents break that assumption entirely, and no regulator is close to filling the gap.

The fragmentation of competing commerce protocols — AP2, Stripe’s ACP, Shopify’s UCP — makes the problem worse, not better. A neutral verification layer sounds like an obvious solution. The harder question is whether the industry will prioritise it before the first systemic failure forces the issue.

REGIONAL

Indonesia plans to embed AI in its US$1.5B free meal programme: The government intends to integrate AI across key public programmes, including its flagship nutrition initiative, signalling a shift toward AI-driven public service delivery at scale.

Sea Limited and OpenAI to expand AI access on Shopee: The partnership will bring OpenAI’s capabilities to Shopee’s users and sellers across Southeast Asia, marking one of the region’s most significant e-commerce AI tie-ups to date.

Indonesia orders Shopee, TikTok Shop, Lazada to cut fees: Jakarta has directed the region’s biggest e-commerce platforms to reduce seller fees, a regulatory move that could reshape margins and competitive dynamics across Southeast Asia’s largest market.

MDEC names Ganesh Kumar Bangah as non-executive chairman: Malaysia’s digital economy agency has appointed the industry veteran to lead its board, a significant governance move as Malaysia accelerates its push to become the region’s digital hub.

Singapore AI inspection startup H3 Zoom raises US$3.6M: The funding will support expansion of H3 Zoom’s AI-powered visual inspection technology, which targets manufacturing and infrastructure sectors across the region.

ChemT nets US$4M to ease cell therapy manufacturing: Singapore-based ChemT raised US$4M to simplify the production of cell therapies, addressing a critical bottleneck in the commercialisation of next-generation medical treatments.

NewGen doubles down on K25 AI livestreaming platform: The company is pushing ahead with a commercial launch of its Asia-focused AI livestreaming platform, targeting the region’s fast-growing creator economy and live commerce market.

FileAI closes funding round to scale document intelligence: Singapore-based FileAI secured fresh capital to expand its AI-powered document processing platform, which automates back-office workflows for enterprises across Southeast Asia.

WeRide partners with Geely to bring robotaxi to Hong Kong: The deal marks a significant step for autonomous mobility in the region, combining WeRide’s self-driving software with Geely’s vehicle manufacturing scale.

---

INTERVIEWS & FEATURES

Agentic commerce’s dirty secret: product data is often wrong: The core problem undermining AI-driven purchasing agents is inaccurate or incomplete product data, which causes errors at scale when AI makes buying decisions autonomously.

15 Thai AI companies betting on products, not hype: A roundup of Thailand’s emerging AI builders reveals a growing cohort of startups focused on domain-specific products in healthcare, logistics, and finance rather than foundational model development.

BioArk: building Asia’s life sciences infrastructure: An in-depth profile of BioArk’s strategy to become the backbone of biotech and life sciences manufacturing and logistics across the Asia-Pacific region.

---

INTERNATIONAL

Groq confirms US$650M raise after Nvidia’s US$20B non-deal: AI chipmaker Groq confirmed the fundraise and said it is re-staffing after Nvidia’s reported acqui-hire attempt collapsed, underscoring the fierce competition for AI inference infrastructure.

Inside Zepto’s profit push ahead of IPO: The Indian quick-commerce firm is aggressively restructuring its unit economics to prove profitability before listing, a playbook that carries clear lessons for SEA’s own quick-commerce players.

Meta taps CRED founder Kunal Shah for WhatsApp, invests US$900M: Meta has appointed India’s Kunal Shah as WhatsApp’s new chief and poured US$900M into CRED, a dual move that deepens Meta’s strategic bet on the South and Southeast Asian market.

Nobel laureate John Jumper leaves DeepMind for Anthropic: The departure of the AlphaFold architect signals an intensifying talent war among frontier AI labs, with direct implications for biotech and AI research investment across Asia.

Trump crackdown on Anthropic: who benefits?: An analysis of how US regulatory pressure on Anthropic could accelerate the rise of rival AI labs and open doors for non-US AI providers in markets like Southeast Asia.

Tech layoffs in 2026: AI cited as leading cause: A running tracker of major global tech layoffs this year shows AI automation as the dominant rationale, a trend with growing workforce implications for SEA’s tech sector.

Anthropic says Claude may want to see your ID: The revelation that Claude could request identity verification raises significant questions about AI trust frameworks, consent, and data privacy standards globally and in SEA.

OpenAI launches initiative to patch open-source bugs: The new programme aims to identify and fix security vulnerabilities in widely used open-source software, a move that could benefit the broader developer ecosystem in SEA.

Ubisoft co-founder Claude Guillemot dies in plane crash: The death of one of the gaming industry’s founding figures marks a significant loss for the global tech and entertainment community.

Shareholders sue Uber’s board over sexual assault incidents: A lawsuit targeting Uber’s board over its handling of safety incidents raises governance accountability questions relevant to platform companies operating across Southeast Asia.

---

CYBERSECURITY

After a bank cyberattack, restoring the wrong data is the real risk: A sharp analysis of post-breach recovery failures argues that corrupted data restoration poses a greater threat to financial institutions than the initial attack itself.

Unpatchable flaw in Apple chips opens door to iPhone jailbreak: Researchers have identified a hardware-level vulnerability in Apple silicon that cannot be fixed via software update, exposing millions of devices, including those widely used across SEA, to potential exploits.

WazirX bets on AI futures trading after US$235M hack: The embattled Indian crypto exchange is pivoting to AI-driven futures products as part of its comeback strategy following one of Asia’s largest crypto security breaches.

Why cyber risk ownership is SEA’s biggest leadership blind spot: Leaders across the region continue to delegate cybersecurity to IT teams rather than treating it as a board-level strategic concern, leaving organisations structurally exposed.

---

SEMICONDUCTOR

Qualcomm nears deal for AI chip startup Modular: Qualcomm is close to acquiring Modular, a move that would bolster its AI inference capabilities and intensify competition with Nvidia and AMD in the on-device AI chip market.

Samsung unveils industry’s fastest UFS 5.0 storage solution: The new UFS 5.0 chip delivers double the speed of its predecessor and is designed to power next-generation on-device AI applications across smartphones and edge devices.

Alibaba chip unit raises registered capital by US$148M: The capital injection into Alibaba’s semiconductor arm signals a renewed push to build homegrown chip capabilities amid sustained US export restrictions on advanced technology to China.

SpaceX’s Colossus data centre raises reflection concerns: Elon Musk’s AI data centre is drawing scrutiny over its environmental and operational footprint, a timely reference point as SEA governments approve large-scale AI infrastructure investments.

---

AI

The AI divide in the Philippines started before AI: The piece argues that structural inequalities in digital access and education mean the Philippines risks amplifying existing gaps rather than closing them through AI adoption.

AI agents are joining the workforce; inclusion must follow: As agentic AI becomes embedded in enterprise workflows, technologists are calling for diversity and inclusion principles to be built into AI agent design from the outset.

SEA’s AI momentum outpaces its institutional maturity: A sobering assessment finds that Southeast Asia’s rapid AI adoption is running ahead of the governance frameworks, talent pipelines, and infrastructure needed to sustain it.

Singapore’s AI opportunity is now about discipline, not adoption: The city-state has moved pastthe question of whether to adopt AI and must now focus on building the organisational rigour to deploy it effectively and responsibly.

---

THOUGHT LEADERSHIP

VC liked you; that’s not the same as yes: A candid examination of how founders misread investor signals during fundraising, confusing positive engagement for commitment, a common and costly mistake in the SEA startup circuit.

When execution is free, the brief becomes the product: As AI commoditises delivery, strategy and clarity of thinking become the scarcest and most valuable inputs, a fundamental shift in how founders and operators should think about their roles.

The next startup opportunities are forming around control: The argument is that as AI automates efficiency gains, the next wave of valuable startups will be those that give users and organisations meaningful control over automated systems.

How AI stocks are stealing billions from crypto: As institutional capital rotates from crypto into AI equities, the piece examines what this structural shift means for crypto valuations and the investor appetite for digital assets in SEA.

Why tracking Bitcoin ETFs matters for SEA investors: Bitcoin ETF flows are becoming a reliable proxy for institutional sentiment toward crypto, offering SEA investors a clearer signal amid market volatility.

Social impact funding needs a common language, not more capital: The piece contends that impact investing in SEA is held back less by a lack of funds than by the absence of shared metrics and definitions across funders and founders.

The Eisenhower Matrix, Maslow, and the goals you set yourself: A reflective essay challenges founders and operators to question whether their goal-setting frameworks serve genuine priorities or simply replicate conventional ambition.

The post Ecosystem Roundup: The agentic commerce trust gap no one wants to fix appeared first on e27.

Singapore has set an ambitious target: 100,000 “AI-bilingual” workers by 2029. The goal signals a broader reckoning with how AI is reshaping the professional workforce — not merely as a productivity tool, but as a capability that demands a new kind of literacy. Yet as training programmes multiply and certification frameworks take shape, a harder question is emerging: what does AI bilingualism actually require in practice?

“AI bilingualism means having enough domain expertise and AI fluency to actually direct, evaluate, and push back on what AI gives you,” says Felicia Tan, Director of Tribe Academy, in an email interview with e27.

The distinction matters. Faster, more polished output is already well within reach for most professionals. The ability to spot where that output is wrong — or quietly dangerous — is proving far more elusive.

Tribe Academy offers expert-led training in areas including AI and blockchain to further bridge Singapore’s talent gap. In our conversation, Tan reveals the blockers that many corporations in Singapore face in embracing AI–and what to do about it.

The following is an edited excerpt of the conversation.

Also Read: 5 Seoul startups made their Southeast Asia debut at Echelon Singapore 2026 under the SBA pavilion

MOM’s latest survey shows 70 per cent of Singapore companies still have not adopted AI for work, a striking number given how much policy attention has gone into this space. In your experience working with corporate clients, what’s the real blocker?

If you spend a lot of time in tech circles, it can feel like everyone is already using AI. But outside that bubble, many organisations are still at the stage of observing, experimenting cautiously, or waiting to see clearer proof of value before changing how work gets done.

Everett Rogers gave us the Diffusion of Innovations curve decades ago, and it remains one of the most useful lenses for moments like this. The theory has long shown that every major shift moves through stages. Innovators, early adopters, early majority, late majority, then laggards, each arriving on their own schedule, for their own reasons. Right now, AI still sits heavily between the early adopterand early majority phase for many Singapore companies.

From the perspective of early adopters, it can feel like progress is slow. But we also need to recognise the scale of behavioural change being asked of the workforce. The oldest members of our working population in Singapore today entered their careers roughly 30 years ago, in the mid-1990s … Entire careers were built around ways of working that rewarded precision, hierarchy, and predictability.

AI changes not just the tools people use, but the nature of how work gets done. That transition naturally takes time, especially at workforce scale. Policies and national initiatives help create momentum, but cultural and operational change inside organisations has always moved slower than headlines.

One main blocker we are seeing with AI adoption is that it is still highly siloed and deeply individual. We see individuals attending our programmes who bring these skills back, but only to their personal chat windows. Someone on the team discovers a prompt that saves them two hours a week, and they quietly use it, and nobody else knows. Someone in HR uses a new AI tool for meeting summaries but the knowledge stays private. There is no institutional memory layer or shared playbook that captures what’s working.

Also Read: The great rotation: How AI stocks are stealing billions from crypto

So you get a patchwork where a few power users produce impressive outputs, while everyone else is doing things roughly the way they always have. The gain will live and die with the individual. For organisation-wide impact, a deliberate redesign of workflows, KPIs, or operating models will need to follow either through top-down directives or a conscientious effort by the entire staff.

The next blocker is arguably the most honest one, i.e. if it isn’t broken, why fix it? Not everyone is a productivity advocate and lies awake worrying about workflow inefficiency. Sure, some firms are redesigning roles and creating new AI-related positions. But AI is not visibly taking anyone’s job tomorrow. Then reasonably, as human beings, we tend to stay with what works, that is, no change is needed.

The third blocker is structural among those who have attempted to look into AI implementation. The most commonly cited constraints are high implementation costs and lack of in-house expertise. The tools exist. The willingness, in many cases, exists too. But the bridge between “I’ve heard of AI” and “We’ve redesigned our workflow around it” is still too long and too expensive for most SMEs to cross without support.

There’s a tendency among early movers to look at policy timelines and grow impatient. But policy takes time to land. The government has announced a new Tripartite Jobs Council to support employers and employees in AI adoption, alongside access to free premium AI tools for Singaporeans taking selected AI courses.

The reality is that a Budget announcement in February does not transform workforce behaviour by April. There will always be a lag between national intent and organisational habit change. The real test is what companies do in that gap. Grants and national initiatives can reduce the risk of taking the first step, but they cannot redesign workflows on behalf of every employer.

Also Read: Agentic commerce’s dirty secret: the data powering AI purchases is often wrong

There’s been a lot of hiring around “prompt engineers” over the past two years. Is that the right unit of skill to be built for? What’s the capability that actually drives business value that most job descriptions and course catalogues are still missing?

Good prompt engineering is underrated. It was the first core skill that emerged when Generative AI became accessible to the general public, and it remains foundational.

The fact that leading AI companies are still publishing prompt engineering guides and running 101 courses around it tells us something. The practical case for why it matters more is important as models get more powerful. The newer reasoning models consume significantly more tokens, especially when you are building complex workflows or automating multi-step processes. If you do not know how to construct a tight, well-structured prompt, you will burn through credits at an alarmingly fast rate.

Extrapolating this across a team running dozens of automated workflows, and it becomes economically untenable. It is a boring skill compared to flashy AI apps and dashboards, but knowing how to communicate with AI systems precisely will save companies enormous amounts of money and frustration over time.

Prompt engineering remains a useful skill to develop, but its real value is as a foundation for broader AI capability, not as the end goal. Prompt engineering gets one to a useful first draft. What you do with that draft … is one that most job descriptions and course catalogues are still fumbling to articulate, but what is going to derive the most business value.

So, if I were advising an enterprise on what capability to actually build for, it would be this: Workers who have developed strong prompt discipline as a baseline habit, and who pair it with critical thinking to know when the model is leading them somewhere wrong. The future is probably less “prompt engineer” and more “AI-native operator”.

Also Read: Why Cyber Risk Ownership Is Southeast Asia’s Biggest Leadership Blind Spot

If you could redesign one thing about how Singapore is approaching workforce AI upskilling right now, what would it be?

If I could redesign one thing about how Singapore is approaching workforce AI upskilling right now, I would shift the focus from primarily funding structured training to creating a much stronger bridge between training and rapid on-the-job application.

Many companies still see upskilling as “time away from real work”. To close this gap, we need structured training to remain the foundation in building mental models and tool confidence, but we must also ensure it then quickly flows into real application. Specifically, companies should set aside dedicated hours each month for employees to test AI on actual tasks, just like how R&D time is ringfenced in tech companies and “timetabled time” is set aside for teachers to dedicate time to innovation and professional development.

Policy can reinforce this by tying enhanced grants to organisations that implement and report on these pilots, with even stronger support when they become sustained initiatives rather than one-off efforts. Once organisations have a core group of upskilled champion users, they should guide the rest of the team to start small with specific tasks, such as shortlisting documents, summarising meeting notes, or automating a two-part workflow.

The goal is to learn by doing something real and low-stakes. Have employees treat early failures as cheap tuition. Just like in the early days of the internet, nobody expected the first company website to generate revenue immediately.

This approach aligns with one of Singapore’s strongest policy philosophies: reducing the downside ofexperimentation. Our grants, co-funding, and training subsidies were never designed to guarantee perfect outcomes, they exist to make the first step less risky and encourage early action.

By redesigning the system this way, we can turn awareness and training into genuine productivity gains and keep Singapore’s workforce truly competitive.

—

Image Credit: Cash Macanaya on Unsplash

The post Tribe Academy’s Felicia Tan: Why good prompt engineering and critical thinking are keys to AI bilingualism appeared first on e27.

H3 Zoom, a Singapore‑based deeptech startup building AI‑driven inspection and asset‑intelligence software, has closed an oversubscribed Series A round of US$3.6 million.

The financing was led by JRE Ventures, the corporate venture arm of Japan’s East Japan Railway Company, with participation from SGInnovate, M7 Holdings, Moringa Ventures, and Lotus One Investment, besides an AngelCentral member syndicate.

Also Read: H3 Zoom lands US$1.8M to accelerate AI-powered building inspections in Japan, HK

The cash will bankroll H3 Zoom’s push across Asia, with explicit emphasis on Japan, Hong Kong SAR, Singapore, and Southeast Asia. The company says it will invest in product development, engineering hires, enterprise go‑to‑market execution and integrations across building and infrastructure lifecycles.

Why investors are paying attention

H3 Zoom has spent the past decade combining computer vision, proprietary vision‑language models, drones and robotics‑assisted capture into a single inspection workflow. The result is a platform that converts photographic and sensor inputs into structured, standards‑aligned reports and analytics, a shift from manual, fragmented inspections toward traceable, repeatable processes.

Investors tell a similar story: infrastructure owners in Asia face an ageing stock of assets, tighter budgets, and skills shortages, creating a market for scalable inspection technologies. For corporate strategic investors such as JRE Ventures, the appeal is both defensive and strategic, ensuring safer operations across rail, station and commercial assets while opening commercial ties across Asian markets.

“Through this investment, we aim to accelerate H3 Zoom’s business expansion and proof‑of‑concept activities in the Japanese market, while exploring broader collaboration opportunities across Southeast Asia,” said Junichi Eto, Managing Director at JRE Ventures. His comment highlights the importance of Japan as a commercial beachhead for the company, and the potential for tech transfer into regional partners and operators.

What H3 Zoom actually sells

H3 Zoom’s headline products –Façade Inspector and Interior Inspector — are focused on reducing inspection time, cutting operational and access costs, and lowering reliance on labour‑intensive work‑at‑height activities. Using drones for data capture, AI for defect analytics and standardised reporting workflows, the company says it helps customers adhere to regulatory frameworks such as Singapore’s Building and Construction Authority periodic façade inspection regime.

Also Read: Transforming asset inspections: How WaveScan’s smart sensors and AI are shaping predictive maintenance

In practical terms, that matters for Southeast Asia. Cities across the region are racing to upgrade ageing building stocks and transport infrastructure while facing tightening labour markets. Local authorities and facility owners increasingly demand verifiable inspection records, and insurers are looking for standardised evidence of maintenance. That creates a commercial runway for software that not only detects defects but also ties findings into asset management systems and maintenance workflows.

Regional traction and repeat business

Investors pointed to H3 Zoom’s customer traction and repeat business as reasons to double down. AngelCentral’s member‑led syndication that topped up the round after an earlier first close was singled out as an important validation of the company’s regional momentum.

“I was not only impressed by the concept, but most of all by the traction the company had already,” said Marnix Beugel, the AngelCentral syndicate lead. The remark underscores a common investor filter in Southeast Asia: demonstrable, recurring revenues from repeat customers often outweigh speculative product roadmaps.

Product roadmap: AI co‑pilot and multimodal workflows

With fresh capital, H3 Zoom plans to accelerate an “AI Engineering Co‑Pilot” and multimodal inspection features combining 360‑degree imagery and voice notes, alongside enterprise‑grade APIs and robotics‑assisted capture. The aim is to make inspections faster and to surface actionable issues for engineers more consistently, turning inspection outputs into measurable maintenance outcomes.

Shaun Koo, H3 Zoom’s founder and CEO, framed the funding as a validation of the company’s mission. “With this capital, we will accelerate our AI roadmap, deepen enterprise integrations, and scale across key Asian markets where infrastructure safety, asset resilience and inspection productivity are becoming increasingly important,” he said.

Competition and the wider market

H3 Zoom operates in a crowded but fragmented space. Startups, system integrators and established engineering firms are all experimenting with drone capture, AI analytics and robotic inspection. Where H3 Zoom hopes to differentiate is through integration: combining capture hardware, proprietary AI models and enterprise workflows that align with regulatory standards.

Also Read: How Japan can empower a new wave of SEA startup innovation

For Southeast Asian operators, the practical considerations are often interoperability and ease of deployment. Systems that bolt on to existing asset management processes and can deliver immediate compliance documentation will likely win the earliest deployments. H3 Zoom’s focus on standards‑aligned reporting in Singapore is therefore a strategic proof point for expansion into nearby markets like Malaysia, Indonesia and the Philippines.

The outlook

The Series A puts H3 Zoom in a stronger position to pursue contracts with asset owners, facility managers and public agencies across Asia. With backing from a mix of strategic (JR East), public deeptech investor (SGInnovate) and regional VCs, the company gains not only capital but commercial channels into Japan and Southeast Asia.

As infrastructure in the region ages and labour costs rise, demand for verification, traceability and decision‑grade inspection data is unlikely to fall. The question for H3 Zoom will be whether it can convert its product depth and early traction into scaled enterprise contracts, and whether its integrations and APIs make it the default “operating layer” for inspection intelligence across Asia‑Pacific.

If it succeeds, the result could be less about replacing humans than about making inspections safer, faster and more auditable — an outcome that resonates as much with regulators and insurers as with engineers on the ground.

The post From façades to railways: H3 Zoom raises US$3.6M to commercialise AI inspection tech across SEA, Japan appeared first on e27.