Posted on — Leave a comment

Everful raises US$48.5M Series A to turn the overseas sourcing office into software

Everful co-founders Kevin Han and Sisi Song

For decades, buying straight from a Chinese factory came with a hidden cost: a sourcing office in Shenzhen or Yiwu. That meant a few agents who spoke the language, chased samples, and knew which supplier was quietly cutting corners. Small retailers who could not afford that office bought through middlemen instead, and paid for it in margins.

Everful, a San Francisco- and Singapore-based startup founded in 2024, is betting that AI agents can do that job. The firm, which builds AI infrastructure for global commerce, has raised a US$48.5 million Series A co-led by Vertex Ventures Southeast Asia & India and January Capital.

Existing investors Square Peg and Lightspeed also joined.

Also Read: Manila has a Series B boom and a Series A funeral. Mynt’s record IPO won’t fix that

The numbers it is touting are large for a company barely two and a half years old. Everful says it ships more than five million pieces a month to retailers and brands in over 60 countries. It says it works with more than “30,000 vetted factories, lists over 20 million items, and is on course to hit US$100 million in annualised revenue by year-end”. None of these figures has been independently verified.

From jewellery wholesaler to 20-million-item catalogue

Everful did not start life as an AI company. It began as a jewellery wholesale business before widening into a multi-category, factory-direct sourcing platform. Co-founder and CEO Sisi Song is a former Alibaba Group executive. She later invested at Bessemer Venture Partners, where much of her focus was cross-border sourcing. Co-founder Kevin Han is an AI entrepreneur who sold his previous startup, Orbeus, to Amazon.

The pitch rests on a familiar pain. Retailers buy through layers of intermediaries and haggle over endless email threads. They also commit to high minimum order quantities months before they know whether a product will sell. If they get the trend wrong, the cash sits in a warehouse as dead stock.

Everful’s answer is what it calls a “world model” of supply and demand. In practice, this is a constantly updated map of what customers want, what factories can make, and how to move goods between the two. AI agents use it to spot trending products, find and vet suppliers, and handle the back-and-forth on every order. Each factory is scored on data from Everful’s own inspections. Customers can buy wholesale, dropship or customise with low minimums, and Everful handles fulfilment end to end.

“Sourcing used to be a people problem. If you were big enough, you hired a team overseas to do it,” Song said. “AI lets us turn that team into software.”

Han frames the difference as outcomes versus tools. “We don’t build tools that leave the work to the customer. We build agents that deliver the result,” he said. He added that every order, inspection and delivery feeds back into the model.

The arithmetic behind the AI branding

“World model” is a phrase now doing heavy lifting in AI startup pitches. Strip it away and Everful looks like a sourcing agent with very good software. That is not an insult. Sourcing has historically been labour-heavy, opaque and stubbornly hard to scale.

Still, the headline numbers deserve a closer look. A rough calculation from Everful’s own figures puts average revenue at about US$1.70 per piece shipped. That points to low-ticket goods, which fits its jewellery origins. Everful takes responsibility for the goods rather than simply matching buyers and sellers. So its revenue probably reflects the full value of what it sells, not a marketplace commission. A US$100 million run rate on that basis is a very different thing from US$100 million in software revenue.

Also Read: The hidden economics of returns in Southeast Asian e-commerce

Geopolitics is the other overhang. Everful sources largely from Chinese factories, and its biggest market is the US. There, the suspension of the de minimis exemption ended duty-free entry for low-value parcels.

Song argues the shift could work in Everful’s favour. She told TechCrunch that long-term cross-border commerce “will be cross-border B2B sourcing plus local selling.” Her logic is that as parcels shipped straight from China to consumers get pricier, local retailers buying in bulk start to look smarter.

Alibaba and Faire have a head start on the data

The B2B sourcing market is crowded. Alibaba.com has long connected global buyers with manufacturers and has rolled out its own AI sourcing assistant, Accio. Faire lets independent retailers buy wholesale directly from brands. Countless sourcing agents and dropshipping outfits sell inspection and logistics as separate services.

Closer to home, Graas is building retail AI agents for brands. Everful’s claim is that it owns the whole chain, from sourcing and order consolidation to quality control and logistics.

That differentiation holds only if the data advantage its investors describe actually compounds. “Their world model of commerce gets sharper with every order,” said Alex Rankin, Partner at January Capital. The same could be said of Alibaba, which has a few more orders to learn from.

Why a Southeast Asia and India fund is leading a US-first deal

The lead investor is a Singapore-headquartered fund, yet Everful calls the US its largest market. The new money will deepen its presence there and in Europe, where the UK, France and Italy are reportedly in focus. Everful will also spend on AI to automate product customisation and on e-commerce integrations beyond Shopify.

The Singapore base explains part of the fit. A broader shift in regional VC explains the rest. Capital has been moving away from mega-rounds towards AI and SaaS, and investors want applied AI companies with real revenue. Everful offers revenue today and an AI story for tomorrow.

There is also a regional opportunity the announcement skips. Southeast Asia’s small online sellers on Shopee, Lazada and TikTok Shop face the same sourcing headaches as American boutiques.

Meanwhile, the region is becoming a manufacturing alternative as investment shifts out of China. A sourcing layer that could route orders to Vietnamese or Indonesian factories as easily as Chinese ones would be a far stronger hedge against the next tariff shock.

Also Read: Southeast Asian tech leaders are learning to trust AI agents, but not with production

Vertex’s Jessica Koh put the case in SME terms: small retailers struggle to find suppliers, “and struggle even more to find ones they can trust.” That trust depends on clean data flowing both ways. Many small merchants in the region still run on patchy records, not tidy dashboards.

The real test will come when things break: a delayed container, a factory that fudges an inspection, or a tariff announced overnight. If Everful can make a boutique in Ohio or a TikTok Shop seller in Surabaya feel like it has a team in Guangzhou, US$48.5 million will look cheap. If the agents stumble, customers will learn what sourcing offices always knew. The hard part was never finding the factory. It was trusting it.

The post Everful raises US$48.5M Series A to turn the overseas sourcing office into software appeared first on e27.

Posted on — Leave a comment

The quiet identity crisis behind AI productivity gains

My work in organisational psychology and clinical practice has increasingly brought me into conversations about how AI is changing people’s sense of professional identity.

One client, whose identifying details have been changed to protect confidentiality, leads a research team at a Singapore fintech company. AI has transformed how she works. Reports that once required hours of analysis can now be produced in minutes, leaving her more time to review findings, interpret results and make decisions.

Her productivity has improved considerably, but the nature of her contribution has changed. Much of the work through which she previously developed and demonstrated her expertise is now completed by AI.

Her experience raises an interesting psychological question: when the tasks that helped establish our professional identity begin to disappear, how does that change our sense of competence and achievement?

What the productivity figures leave out

According to Microsoft’s 2026 Work Trend Index, 66 per cent of AI users in Singapore report producing work they could not have created a year earlier, compared with 58 per cent globally.

For businesses looking to scale, these are impressive gains. Figures like these make their way into boardroom discussions and strategic planning. Few organisations are examining what happens to employees’ sense of competence as the nature of their contribution changes.

A 2025 study published in Scientific Reports, involving 3,562 participants across four experiments, found that working with generative AI improved immediate task performance. However, participants who subsequently returned to independent tasks reported lower intrinsic motivation and greater boredom.

Interestingly, their sense of control increased when they resumed working independently.

We develop our professional identity partly through effortful mastery. Solving difficult problems, making decisions and seeing the results of our own thinking provide evidence of what we are capable of.

When AI begins performing tasks through which people previously demonstrated those abilities, the experience of achievement can change, even when the quality of their work improves.

Also Read: Is prompt engineering real AI development, or just good documentation?

When confidence begins to erode

Research from Microsoft and Carnegie Mellon University surveyed 319 knowledge workers who regularly used generative AI. Greater confidence in AI was associated with less reported critical thinking effort, while confidence in one’s own abilities was associated with more.

The findings don’t establish that AI causes skills to deteriorate. They do, however, raise questions about what happens when people have fewer opportunities to exercise their abilities independently.

In psychology, self-efficacy describes our belief in our ability to perform a particular task. That belief develops partly through experience. We become confident in our abilities because we have repeatedly demonstrated them to ourselves.

Someone who rarely completes a complex assignment independently may gradually become less certain they can still do it. That uncertainty can encourage further reliance on AI, leaving fewer opportunities to practise.

Over time, this could create a cycle in which increasing dependence on technology makes independent performance feel more difficult, even as overall productivity continues to improve.

When success no longer feels like your own

The imposter phenomenon traditionally describes capable people who struggle to internalise their achievements, often attributing success to luck or external circumstances.

AI introduces a related question about how people understand their contribution to successful outcomes.

Someone might produce excellent work with AI assistance while becoming less certain which parts of the process reflect their own expertise. If the technology takes over much of the initial analysis, writing or problem-solving, the experience of achievement may also change.

Also Read: What’s actually changing for Asia Pacific businesses in the age of AI

The psychological implications are worth exploring. Professional identity develops through repeated experiences of competence, recognition and personal accomplishment. When those experiences change, people may need to reconsider how they understand their contribution.

From a clinical perspective, I am interested in how people maintain a sense of self-efficacy when the skills they have spent years developing are being used differently.

Rather than assuming that reassurance will address these concerns, there is value in exploring which aspects of someone’s work still depend on their expertise, where their judgement adds value and how they can continue developing their abilities.

In Microsoft’s Singapore findings, 52 per cent of respondents identified critical thinking as the most important human skill for an AI-enabled workplace. Employees appear to recognise the importance of retaining the capabilities that allow them to evaluate, challenge and improve what technology produces.

The difficulty comes when organisations celebrate those capabilities while designing work that gives people fewer opportunities to use them.

What leaders can do

For founders and business leaders, this presents a challenge alongside productivity targets. As AI changes how work gets done, organisations need to consider how people continue developing a sense of mastery and professional value.

  • Protect some independent thinking. Not every task needs to be completed at maximum speed. Teams can deliberately preserve opportunities for employees to analyse information, develop arguments or solve problems without AI assistance. Regular independent practice helps people maintain and develop their abilities.
  • Make human judgement visible. When AI produces much of the initial output, recognising performance becomes more complicated. Leaders should acknowledge the decisions employees make, the errors they catch and the improvements they introduce. These contributions can be substantial, even when they are less obvious than the volume of work produced.
  • Talk openly about changing roles. Only 24 per cent of Singapore respondents in Microsoft’s research reported clear and consistent leadership alignment on AI. When leaders remain vague about what technological change means for employees, people are left to draw their own conclusions. Be clear about what expertise looks like in an AI-enabled role, which skills remain essential and how professional development will continue.

For startups, where responsibilities can evolve rapidly as businesses grow, these conversations are particularly relevant. The pressure to move faster should leave room for people to develop alongside the technology.

Hours saved, and what fills them

AI offers organisations substantial opportunities to improve performance, and many employees are already benefiting from the time and resources it saves.

Yet as technology changes the process of working, it can also influence how people experience accomplishment. Autonomy, mastery and the ability to recognise our own progress contribute to sustained motivation and engagement.

These are aspects of working life that deserve consideration as organisations redesign roles and expectations around AI.

The hours saved are easy to measure. Changes in how people experience their own professional competence are considerably harder to detect.

AI can give teams back hours. Leaders should also consider what those hours are being replaced with, and whether employees still have opportunities to feel proud of what they bring to work.

—

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 quiet identity crisis behind AI productivity gains appeared first on e27.

Posted on — Leave a comment

America’s affordability election is becoming corporate America’s accountability test

When Americans go to the polls on November 3, much of the attention will be on which party controls Congress, but for corporate America, the more consequential question may be what happens after the votes are counted.

Affordability has become a defining issue in the 2026 midterm elections; from groceries and housing to healthcare and electricity, American households remain concerned about the cost of everyday life. An October Pew Research Center survey found that 84 percent of registered voters considered the economy very important to their midterm vote.

Politicians are under pressure to respond, yet businesses are increasingly drawn into the same debate. Companies can point to higher input costs, supply chain disruptions and geopolitical uncertainty to explain rising prices. Many of these pressures are genuine, but consumers and policymakers are asking another question: when companies report strong profits, are customers sharing in the benefits?

The midterms may shape Congress, but the affordability debate will continue beyond Election Day. For American businesses, it is becoming a test of corporate accountability.

From political issue to corporate responsibility

Affordability has traditionally been viewed as a government responsibility. Policymakers manage inflation and regulate markets, while businesses focus on profitability and competitiveness. However, that distinction is becoming harder to sustain.

Penta Group’s latest US policymaker research found that 64 percent identified the economy as their top issue, the first time that figure has exceeded 60 percent in 15 years. Importantly, 56 percent attributed affordability challenges primarily to macroeconomic or geopolitical forces, while 30 percent pointed to companies prioritising profits.

These findings reveal an important distinction. Policymakers recognise that businesses do not control every factor driving higher prices, but they increasingly distinguish between companies that contributed to the problem, those that benefited from it, and those helping consumers manage the consequences.

This scrutiny crosses political lines – Republicans have emphasised making everyday life affordable, while Democrats have questioned whether corporate pricing practices place excessive burdens on consumers. The approaches differ, but both reflect growing expectations that businesses demonstrate how their decisions affect household finances.

When profits and affordability collide

The tension becomes particularly visible when companies report healthy earnings while consumers struggle with essential expenses.

Between September 2020 and September 2026, overall US consumer prices rose by approximately 29 percent, according to Bureau of Labor Statistics data. Although inflation has moderated from earlier peaks, households still face substantially higher prices for everyday goods and services.

Also Read: Why I filed a consumer complaint against 7 global AI giants over the ‘linguistic tax’

Yet the debate is not simply about rising prices; it is increasingly about who benefits when economic conditions improve.

Consider the debate over tariff refunds following the US Supreme Court’s February decision invalidating tariffs imposed under emergency economic powers. Costco pledged to return the value of recovered tariffs to members through lower prices and better value, while FedEx and UPS committed to refunding customers who had paid applicable tariff charges. Other organisations have been non-committal about how customers will benefit from those very refunds.

The controversy raises a broader question: when businesses recover costs previously borne by consumers, how much of that benefit should they pass back?

Companies may have legitimate reasons to retain some gains, including recovering earlier losses or funding investment. But they should be prepared to explain their decisions. The issue is not whether companies should make profits, it is whether commercial success delivers demonstrable value for customers.

Affordability is also about access and fairness

Price alone does not determine whether consumers regard a company as fair. Increasingly, affordability also depends on access, transparency, and whether customers feel they are receiving reasonable value.

Airlines illustrate this tension. Premium cabins, priority boarding, and additional services are legitimate ways to differentiate products and generate revenue, but when basic fares become increasingly restrictive, with extra charges for baggage, seat selection, and other services, consumers may question whether the advertised price reflects the real cost of travel.

The concern is not that premium experiences exist; it is whether businesses are improving choice or gradually making previously standard services available only to those willing to pay more. There is a difference between making premium experiences more attractive and making ordinary experiences less accessible. Across industries, businesses risk reputational damage when pursuing higher-value customers comes at the expense of affordability for everyone else.

Technology introduces another complication. Artificial intelligence (AI) offers opportunities to reduce costs and improve productivity, yet increasingly sophisticated personalised pricing raises questions about whether companies might use consumer data to charge different customers different prices for comparable services.

For policymakers, these practices raise concerns beyond inflation, extending to competition, consumer protection, and corporate governance.

Also Read: The consumer is still human. The gatekeeper isn’t

Proof matters more than promises

The answer is not for companies to sacrifice profitability or absorb every cost increase, but to demonstrate that commercial success and consumer value can go hand in hand. That requires businesses to assess how pricing, fees, and product decisions affect different customer groups, especially those with fewer alternatives, and to identify opportunities to deliver value while sustaining growth. They should also demonstrate tangible benefits. If supply chain improvements reduce costs, are savings passed on? If automation improves efficiency, does it make services more accessible? Are affordable options preserved?

Boardrooms should treat affordability as part of discussions about reputation, regulation, and long-term competitiveness, not simply as a communications challenge. Penta’s policymaker research offers an instructive example. Multiple policymakers across party lines independently identified Costco as an affordability leader. Respondents pointed to its reputation for competitive prices,  relatively low margins and focus on customer experience.

That recognition reflects consistent business practices rather than a single marketing campaign. Credibility comes from what businesses do, not simply what they say.

Beyond November

The implications extend beyond the United States. American debates about corporate pricing, consumer protection, and responsible AI use can shape international regulatory expectations, including among multinationals operating in Asia.

Companies that treat affordability as somebody else’s responsibility risk greater consumer skepticism and regulatory scrutiny. Those that show their decisions preserve access, improve transparency, and deliver value will be better positioned to maintain trust.

Businesses may not be responsible for every increase in the cost of living, but they will increasingly be judged by how they respond to rising affordability pressures. The midterms will determine who holds political power in Washington, but the affordability debate will continue to shape how policymakers, consumers, and other stakeholders assess the way corporate America exercises its economic power.

Co-authored together with Bryan DeAngelis and Tucker Warren, Senior Partners, Penta Group

—

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 America’s affordability election is becoming corporate America’s accountability test appeared first on e27.

Posted on — Leave a comment

AI should change wealth management’s economics, not just its margins

Implementation of Artificial intelligence (AI) across the industry is giving wealth firms a choice in how they use the savings it generates, whether to lower fees, serve smaller accounts, improve access to financial advisers or retain the savings as profit. However, these benefits do not automatically reach clients. Firms must make deliberate decisions about how the savings are used and measure whether they lead to meaningful improvements in affordability, accessibility and service quality.

For decades, personalised wealth management has largely been reserved for those with substantial assets. The economics make most sense at those numbers, as private banking is a high-touch, relationship-driven model that requires a certain level of cost. It follows that investors with smaller portfolios have typically been offered standardised products and digital platforms, catered for broad risk profiles rather than the ongoing advice and personal attention available to wealthier clients.

This gap is a huge opportunity for an industry that has traditionally struggled to serve smaller investors profitably. AI changes this equation and challenges the assumption that service quality must be tied to portfolio size. 

AI changes this equation and challenges the assumption that service quality must be tied to portfolio size.

How AI can free up advisers to focus on clients

Much of the conversation around AI in wealth management has focused on productivity. And AI does this well: automating administrative tasks, investment reports, and client onboarding. While these improvements address a persistent challenge in Singapore’s wealth management industry, where outdated systems and manual processes can limit the time advisers spend with clients, companies need to consider where and how these changes will make the most impact in their client-servicing.

Consider onboarding a new client. Collecting documents, verifying identities, conducting risk assessments and preparing suitability reports all can be done by AI in much less time and manpower. The same applies to portfolio reviews: gathering investment data, checking portfolio performance, identifying changes and preparing reports all in a few clicks.

Meanwhile, wealth advisors now can focus on governance and on understanding their clients’ unique needs while establishing a deeper servicing relationship.

Also Read: Before you add AI to your product, fix how work moves inside the company

Making personalised wealth management viable for smaller investors

The time saved from AI implementation is a gross figure. Integration, data preparation, output checks, exceptions and ongoing oversight all carry costs, and suitability and compliance work does not disappear. The Monetary Authority of Singapore’s guidelines on AI risk management, issued on 7 October 2026 and taking effect from 7 October 2027, set expectations on accountability and human oversight. Those governance costs belong in the business model.

For smaller accounts, the regular costs like time and resources for meetings, portfolio reviews, compliance and communication may outweigh the revenue generated, leading many wealth firms to set minimum investment requirements based on their service offerings.

An investor with US$200,000 may not require the same volume of service as someone managing US$5 million, they both deserve the same quality of personalised attention. Portfolio size alone does not establish how complex a client’s needs are in real life, and could be trying to balance liquidity, medium-term goals, and long-term wealth creation simultaneously. 

AI implementation allows personalised investment support to scale because institutional-grade analytical capability becomes possible at a consumer-level cost, while enabling relationship managers to improve their services qualitatively.The risk, however, is that firms use technology to increase client numbers while reducing human support. An investor offered automated notifications and occasional chatbot interactions may have cheaper access to investment products, but not necessarily better advice.

Making wealth management more affordable should not mean making it less personal.

Why human advice remains important in an AI-driven market

According to HSBC’s 2026 Human-AI Advantage study, 76 per cent of Singapore’s mass-affluent and high-net-worth investors use AI for financial purposes, but many still seek an adviser’s reassurance before acting. This is happening across generations, not just the young. 

Ultimately, clients are looking for something that AI cannot give. Human advisors can fill in the gap, providing trust and cater hyper-personalised advice not only based on the numbers, but also the client’s concerns and qualitative needs. We’re dealing with clients who have real financial complexity, and who want to know how to balance liquidity, medium-term goals, and long-term wealth creation all at once. AI cannot provide this.

Also Read: Cheques, bans and bell-ringing: 50 Asian AI companies that defined 2026

However, with easier access to investment research and financial tools, investors can arrive at meetings better informed and with more specific questions. Advisers must therefore demonstrate value beyond providing information, focusing instead on interpreting findings, challenging assumptions and helping clients make decisions suited to their circumstances.

Measuring whether AI savings benefit investors

As AI improves efficiency, wealth firms must consider where the savings ultimately go.

Firms need healthy margins to operate sustainably. However, operational efficiency should not automatically be presented as a customer benefit.

A firm that doubles the number of clients each adviser manages may improve its profitability. But if clients receive slower responses, fewer reviews or less personalised guidance, the improvement is largely internal.

Rather than focusing solely on productivity, firms should measure whether AI improves service quality and clients’ access to advice. Key indicators include servicing costs, fees, minimum account requirements, adviser availability, response times and client complaints. 

Digital wealth platforms have made investing more affordable, but access to cheaper products does not necessarily mean better access to personalised advice. AI could help firms offer services based on client needs rather than assets alone, while maintaining appropriate oversight.

Personalised, quality wealth management, provided by experienced investors and expedited by AI models, should naturally include passing savings on to smaller investors. Access to ongoing advice could improve client retention, encourage referrals and reduce acquisition costs as clients’ wealth grows. However, these benefits must be supported by actual retention and cost data. The same considerations apply across the wealth management industry.

The credible AI wealth platform will be able to show not only how much work it removed, but what its clients gained as a result.

—

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 AI should change wealth management’s economics, not just its margins appeared first on e27.

Posted on — Leave a comment

We took founder bias out of the roadmap: Then our customers fired us

The pristine whiteboard mounted on the corner wall of the fifth-floor incubator suite at one-north was completely bare, save for the ghost of an erased product flow from three months back. In the late nineties, covering the early wreckage of the dot-com boom out of San Francisco, you could smell a founder’s panic fifty feet down the corridor: paper taped to drywall, frantic marker scrawls, half-eaten takeaway cartons, and founders arguing with their lead engineers until their voices gave out. Decisions were hammered out through exhausting, imperfect human friction. Here, in the glass-and-steel corridors of Singapore’s technology hub, the suite had the airless calm of an intensive care unit.

Across the pale birch table sat Julian, a twenty-nine-year-old founder whose logistics SaaS platform had closed an oversubscribed seed round fourteen months prior. His laptop was open to a dark terminal where background scripts ticked along in eerie silence. A constellation of autonomous agents ran continuous loops: one scraped regional competitors’ pricing tiers across ten Southeast Asian jurisdictions, another synthesised thousands of customer support logs into thematic clusters, and a third continuously re-ranked their product backlog based on weighted sentiment scores.

Julian ran his thumb along the aluminum edge of his laptop with the quiet pride of someone who believed he had outsmarted entropy. He told me his team had stopped having roadmap arguments because the pipeline ingested the telemetry, scored feature requests against market velocity, and spit out sprint objectives with confidence intervals attached. Every single goal was mathematically defensible. With a contented smile, he leaned back into his ergonomic mesh chair and told me they had successfully eliminated founder bias and engineered out friction. Three weeks later, his lead venture backer refused to write a Series A check, and by the end of the quarter, two of his four flagship enterprise accounts had walked away.

Over a quarter-century of watching technological shifts swallow industries whole, I have seen the same tragedy stage itself across different eras, changing only its vocabulary. In the early 2000s, automated credit-scoring models promised mortgage desks that algorithms could price collateralised debt without the erratic interference of loan officers’ subjective hunches.

In 2008, when the subprime architecture imploded, the executives sitting before congressional committees looked dazed because their models had pronounced systemic risk nearly impossible, yet they had forgotten how to look out the window and evaluate credit with their own eyes. Their critical faculties had atrophied through decades of offloading responsibility to math.

Medicine calls it atrophy when a limb is locked in plaster: the bone mends, but the underlying muscle tissue starves from lack of resistance, shrinking until it can no longer support the body’s weight. Across the venture corridors of Singapore, Jakarta, and Silicon Valley, an identical pathology is taking hold—strategic AI atrophy, the quiet decay of founder conviction and intuitive discernment brought on by outsourcing ambiguous, high-stakes decisions to synthetic consensus. Julian’s seed round had bought world-class engineering, but his automated operational hygiene had quietly amputated his team’s gut.

Also Read: Dubai is selling trust, not just capital: What it means for Southeast Asian founders

What Julian celebrated as eliminating bias was what researchers at Harvard Business School, MIT, and Boston Consulting Group diagnosed with far more clinical precision in Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality.

In studying professionals given access to frontier language models, they documented a dangerous psychological surrender: workers handed smooth, confident algorithmic outputs routinely succumbed to falling asleep at the wheel. Because the machine’s reasoning was synthetically polished and immediately coherent, the human operators stopped interrogating the underlying premises, trading the exhausting labor of critical friction for the effortless glide of automated consensus.

When I asked Julian to walk me through the strategic pivot he had executed that spring, he clicked into an archived project folder. His largest logistics customer in Indonesia—a legacy distribution network operating hundreds of inter-island freight lanes—had repeatedly begged for an awkward, bespoke offline export tool so warehouse dockworkers could manually override automated routing manifests via WhatsApp threads whenever severe monsoon weather grounded small cargo vessels.

Julian explained that the model had flagged the Indonesian request as an anti-pattern burdened with high servicing overhead, low gross margin potential, and bespoke maintenance risk. The competitive synthesis engine had cross-referenced forty public SaaS post-mortems and recommended doubling down on an automated, standardised self-serve dashboard instead, rendering the statistical verdict unequivocal.

When I asked if he had boarded a flight to Jakarta, if he had spent forty-eight hours on the ground at the Tanjung Priok terminal standing in the rain to watch how freight managers route trucks when port gates flood, Julian looked at me with the polite, bemused tolerance technical founders reserve for print journalists. He asked why he would waste three days stuck in Jakarta traffic when an agent could parse ticket logs, run sentiment classification on customer emails, and cross-reference feature sets against regional SaaS benchmarks in under a minute.

He was reciting what tech companies love to call data-driven objectivity. Yet as ethnographer Tricia Wang observed in Why Big Data Needs Thick Data, big data and quantitative telemetry routinely blind companies to critical human shifts because they only measure the variables an algorithm has been instructed to track. In Jakarta, logistics did not run on clean, cloud-native APIs. It ran on personal relationships between dock clerks and independent truckers, sudden cash-on-delivery settlements, and informal workarounds. The clunky custom feature the Indonesian client had begged for was not an operational inefficiency; it was their lifeline in an unpredictable physical environment.

Also Read: If nobody knows you, nobody owes you their trust: How founders can build it anyway

When Julian’s pipeline quietly stripped that feature from the product roadmap, the customer did not file a bug ticket or register a complaint in the system. They simply recognised that the founders in Singapore had no genuine understanding of how commerce operated in the archipelago, cancelled the pilot, absorbed their losses, and went back to spreadsheets.

The seductive hazard of running an enterprise on AI models is that it constantly confuses probability with breakthrough insight. A foundational model is, by its mathematical definition, an engine of historical consensus. In a study published in Science Advances titled Generative AI Enhances Individual Creativity but Narrows Collective Diversity, researchers discovered a striking paradox: while generative AI tools help individuals write cleaner, more persuasive prose faster, the collective pool of ideas produced across entire groups becomes narrow, predictable, and monotonous.

When a founder asks a model to evaluate an early-stage market, prioritise a roadmap, or structure a commercial offering, it searches the statistical centre of the training corpus to select the highest-probability path—merely the average of how hundreds of other software businesses have solved adjacent problems in the past. As computer scientists documented in Nature regarding AI Models Collapse When Trained on Recursively Generated Data, recursive reliance on synthesised data systematically erases long-tail anomalies, which is precisely where unexpected breakthroughs live.

Every transformative technology business in history was born out of an insight that a consensus model would have scored as an anti-pattern. Renting air mattresses on a stranger’s floor was an operational and liability nightmare, selling physical books on the internet in 1995 offered paper-thin retail margins alongside astronomical shipping costs, and manufacturing high-end electric roadsters when lithium cells cost nearly a thousand dollars per kilowatt-hour was an economic impossibility.

Breakthrough ventures succeed precisely because of the eccentric, unquantifiable convictions of human founders who refuse to accept historical averages. When an entire ecosystem of founders feeds its strategic dilemmas into the same few foundational models, the inevitable result is collective homogenisation: everyone builds the same feature sets, adopts the same sanitised landing-page copy, and deploys the same three-tiered pricing strategy. They build companies that are mathematically defensible, structurally sound—and completely devoid of a competitive soul.

As economists Ajay Agrawal, Joshua Gans, and Avi Goldfarb observed in The Economics of Artificial Intelligence, when machine intelligence makes predictions and routine cognitive execution cheap, the economic value of the complementary human asset skyrockets. That complementary asset is not prompt syntax; it is judgment, the willingness to bear the weight of an unhedged decision when historical data offers no clear answer.

Also Read: Build what matters or what gets funded? India’s VCs say the real gap sits in the founder’s head

Two months after our initial conversation, I met Julian again at an open-plan co-working space on Robinson Road. His company had surrendered its leased incubator suite, and the engineering team had been trimmed from 11 to four. His screen was open to an empty text editor, the terminal of autonomous agents quiet for the first time in a year. Stirring a cup of black coffee that had gone cold, he admitted that the hardest part was realising that when the market pushed back, they had no muscles left to push back with.

For 14 months, whenever they hit a crossroads regarding pricing tiers, enterprise retention, or product positioning, they had fed the variables into the agents. They had debated the summaries, but they had never practised the agonising work of wrestling with the ambiguity themselves. They had treated the prompt output as the oracle, and when the enterprise clients left, the agents had no answers, merely generating four more variations of the same generic SaaS playbooks. Staring out across the dense mid-morning traffic flowing toward Shenton Way, Julian whispered that while they thought they were moving faster than everyone else, they had actually been letting the machine think for them until they forgot how to think for themselves.

The lesson for this generation of operators is not a Luddite call to close the terminal or throw away the tools. AI agents are astonishing instruments for automating baseline operational drag, perfectly suited to parse code, clean datasets, transcribe calls, and summarise telemetry. The machine should manage the operational floor, but it must never be allowed to usurp the strategic ceiling.

If you are building an organisation, you must guard against the silent creep of AI atrophy by using models to cluster user tickets while continuing to board the plane, sit across the table from your customer, and watch the unspoken hesitation in their eyes when they talk about their business. You must study consensus outputs specifically to recognise the average, predictable path your competitors will take, using human intuition to chart the contrarian path they cannot see.

Above all, you must protect the muscle of accountability. When a decision carries existential risk, refuse the comforting anaesthesia of an algorithmic score, sit in the quiet, feel the weight of the ambiguity, and make the call yourself. In an economic landscape flooded with infinite, auto-generated certainty, human conviction is the only asset that cannot be commoditised. The algorithm can calculate the odds, but only a founder can accept the consequences. If you surrender your judgement to the machine, you may build a frictionless company, right until the moment you discover you have automated away the only reason you were ever in the room.

—

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 We took founder bias out of the roadmap: Then our customers fired us appeared first on e27.