For two decades, the startup gospel was simple: find product-market fit at home, then blitzscale across borders. In 2026, much of the world appears to be rethinking the second half of that sentence.

According to HSBC’s Global Entrepreneurial Wealth Report (GEWR) 2026, the share of entrepreneurs globally who want to expand internationally fell six percentage points year-on-year to 43 per cent. Interest in domestic expansion and local mergers and acquisitions (M&A) jumped from 42 per cent to 64 per cent, making it the most popular growth strategy worldwide.

Also Read: What I have learned watching foreign tech companies build in Singapore

The backdrop is familiar. Governments want control over data, energy grids that power AI data centres, and critical tech supply chains. When cross-border deals face growing regulatory friction, buying a capability at home looks easier than building one abroad. Technology leads all sectors in domestic M&A appetite, with 69 per cent of tech entrepreneurs pursuing domestic acquisitions, followed by energy and mining (65 per cent), manufacturing (56 per cent) and electronics (54 per cent).

The most dramatic swing is in the US, where domestic acquisition intent reportedly rose from 19 per cent in 2025 to 62 per cent. A threefold jump in a single year says as much about the mood in Washington as it does about business strategy.

The ASEAN exception

Southeast Asia is moving the other way. Indonesia ranks first globally for international expansion intent, with 62 per cent of Indonesian entrepreneurs prioritising overseas growth. Thailand is second at 55 per cent and Malaysia third at 52 per cent, ahead of mainland China at 50 per cent.

This isn’t a new pattern. In last year’s edition, the top five countries where entrepreneurs were interested in international expansion were Indonesia, Malaysia, the UAE, Saudi Arabia and China. The region’s outward lean is becoming a fixture of the series rather than a one-off spike.

HSBC attributes this to regional integration and a consumer boom, projecting that private consumption across the six core ASEAN economies will grow about 8 per cent a year to reach US$5 trillion by 2035.

Before treating this as a pan-ASEAN startup stampede, though, consider who is answering the survey. The GEWR is a private banking study. In its 2025 edition, HSBC surveyed 2,939 HNW business owners with at least $2 million of investable assets; a third of the sample (956) were ultra-HNW individuals with a net worth of more than US$100 million. These are established, wealthy business owners, many of whom run family conglomerates, not seed-stage founders burning through a Series A runway. Their appetite for expansion is a useful signal, but it isn’t a startup census.

There is also a less flattering reading of Indonesia’s top ranking. Outward ambition can reflect confidence, but it can also reflect a wish to diversify away from home-market risk, currency swings and regulatory unpredictability. The survey measures intent; it cannot tell the two motives apart.

Singapore, the waiting room

Singapore, unsurprisingly, emerges as the region’s anchor. According to the report, 40 per cent of entrepreneurs globally are more likely to consider expanding into Singapore over the next 12 months, putting it level with Japan as the world’s top expansion destination.

When entrepreneurs choose where to expand, their priorities are access to advanced technology (30 per cent), a broad customer base (28 per cent), technology-friendly policies (27 per cent), professional services (26 per cent) and skilled labour (25 per cent). Singapore ranks first globally for simple and stable regulation, joint second for innovative tech policy, and third for advanced AI infrastructure. It also tops the list for multi-residency, with 9 per cent of business owners planning a residence there.

Also Read: The cross-border due diligence questions most founders cannot answer

That last figure is a reminder of what kind of report this is. A meaningful part of Singapore’s appeal to this cohort concerns where they park their families and their wealth, not just where they build companies. For a bank whose private banking business depends on precisely those flows, it is a convenient finding.

Singapore’s gravitational pull also carries costs for the rest of the region. The city-state now captures 78 per cent of SEA’s startup funding, and its stability does not make it immune to sovereignty politics. The case of Manus, which went home despite its Singapore base, showed that a Singapore address cannot always shield a company from geopolitical pressure.

The playbook, with the price tags attached

The conventional advice drawn from the data is a “dual-track” model: anchor intellectual property, treasury and fundraising in Singapore, then run localised “hub-and-spoke” operations in Indonesia, Thailand, Malaysia and Vietnam. On paper, it is sensible.

In practice, regional expansion has humbled many who tried it. e27 recently examined why Indonesian fintech’s ASEAN expansion cost more than anyone priced in, and the lessons from Matchmade.io’s attempt to sell Indonesian finance tech in Singapore were similarly sobering. Localisation remains stubbornly manual: AI is changing global expansion, but it cannot standardise local markets.

Data sovereignty is the trap most founders underestimate. Each market’s data residency rules require separate architecture, compliance and sometimes separate entities, yet Singapore firms fear data sovereignty failures while remaining underprepared.

Acquisitions offer a shortcut, since buying a local player brings its licences and distribution with it. Regional consolidators are already doing this: logistics firm Amilo has completed its fourth buyout in four years. But cross-border M&A is only as good as the homework behind it, and many startups still struggle with the due diligence questions most founders cannot answer.

Ambition is not the bottleneck

The GEWR’s headline message, that Southeast Asia is the world’s most outward-looking region while others retreat, is both true and incomplete. Ambition has never been the region’s scarce resource. Execution is: the cost of entering ten regulatory regimes, the talent to run them, and the capital to survive the years it takes.

Also Read: Why Singapore, Indonesia, and Vietnam are losing the AI race they think they are winning

The ASEAN Digital Economy Framework Agreement may lower some barriers, but as e27 has noted, DEFA cuts the cost of entry while architecture sets the cost of scale.

The world’s wealthy may be turning inward, and Southeast Asia’s are heading for the door. The question is how many of them will still be abroad by the time next year’s survey comes around.

The post The world’s rich are turning inward. Southeast Asia’s entrepreneurs are packing their bags appeared first on e27.

Ask a care manager in Singapore when the workday ends and the honest answer is: some time after the last visit, once the roster is patched, the visit notes are filed and the calls to anxious families about a missed shift are done.

Homage, the Singapore-founded home care company, thinks that unglamorous back office is a business in its own right. On Wednesday it launched Esteem AI, care management software for home care, aged care and disability support providers, and set it up as a standalone regional business alongside its care services. A global rollout, the company says, will follow.

Also Read: When a slot opens, let the AI agent act – within limits

The software itself is not new. Homage built it to run its own operations in Singapore, Malaysia and Australia, and more than 50 care organisations already use it, including Home Nursing Foundation (HNF), a charity that has sent nurses into Singaporean homes since 1976. What is new is the name, the separate business line and a layer of AI agents meant to take the paperwork off care managers’ plates.

The timing is no accident. Singapore became a super-aged society this year, with more than one in five citizens aged 65 or older, and the Ministry of Health expects a million seniors by 2030. Every one of them brings care needs, and every care visit brings forms, audits and rosters.

From three tools to one record

Esteem folds client intake, rostering, workforce management, incident reporting, customer support, billing, reporting and compliance into a single system, keeping each client’s history on one record. Homage’s argument is that most agencies still run on a rostering tool, a spreadsheet and a shared inbox that were never designed to talk to one another.

The AI layer is where the launch gets its headline, and also where the fine print matters. Most of the agentic features in the release are described in the future tense. Homage says it is building agents that will draft a visit note from what a care worker recorded, suggest a replacement carer when a shift opens up based on skills, location and whether the client already knows them, chase workers whose certifications are about to lapse, and answer routine questions from families using the record, passing anything that needs judgement to a human.

How much of that runs in production today is not spelt out. What Homage does stress is the guardrail: nothing that affects a client’s care or a worker’s schedule happens without a person’s approval, and every action is logged.

“Care will always be about people,” said Gillian Tee, co-founder and CEO of Homage. “Our AI agents prepare the work, but the decisions stay with care teams, as they always have.”

Also Read: Agentic AI is Singapore’s hottest job skill, and that should worry the people it replaces

That human-in-the-loop framing fits the regional mood. Southeast Asian tech leaders are learning to trust AI agents, but not with production, and regulated sectors want someone clearly accountable when an agent acts. In care, the bigger risk is quieter than a rogue agent: an exhausted care manager approving an AI-drafted note at 10pm without reading it closely. A sign-off button is only a safeguard if people treat it as one.

Six years in the making

Homage has flagged this move for a long time. As early as 2020, it set out plans for platforms other care providers could use. Tee says the push came from inside the company. “What surprised me was that everything around the care was just as hard: the notes, the rosters, the forms, the calls to families,” she said.

The traction figures are respectable rather than spectacular. More than 90 per cent of the 50-plus organisations use Esteem as their everyday care management system, and over 300 care managers use it in an average month, roughly six per organisation. HNF, the marquee customer, runs its home personal care operations on the platform and serves more than 1,000 care recipients a month. Its CEO, Dr Christina Tiong, said the platform is especially valuable outside office hours, helping the charity deliver continuous care “without adding administrative strain”.

Homage did not disclose revenue, pricing, or how many of those organisations pay commercial rates.

There is also an awkward structural question. Homage remains a care provider, and some of the agencies it now courts as software customers compete with it for clients and carers. The release does not say how customer data will be walled off from Homage’s own care business, or whether one agency’s records will help train agents used by another. In a sector handling sensitive health information, those are the first questions a procurement team will ask.

Rivals with a head start

Esteem walks into a crowded market. In Australia, one of Homage’s own markets, Canada-based AlayaCare and homegrown ShiftCare, which claims more than 6,500 care provider businesses as users, already sell all-in-one platforms covering rostering, compliance and billing for aged care and disability providers.

At home, Homage still competes on the services side with players such as Jaga-Me. Its pitch is that Esteem was built by an operator for operators. That is a real edge if the agents deliver, and a thin one if they stay on the roadmap.

A second story for investors

The spin-out also hints at how Homage wants to be valued. Founded in 2016, the company has raised more than US$45 million, including a US$30 million (S$40 million) Series C led by Temasek-owned Sheares Healthcare in 2021, and reportedly reached financial breakeven in early 2025. Software margins look far more attractive than labour-heavy care services, and a standalone Esteem gives Homage a second story to tell.

Also Read: The hidden economics of autonomous AI agents

Its track record on expansion counsels caution. In 2020, Homage said it planned to enter five more countries within two years, and announced a push into Japan with backing from Infocom. Neither has visibly materialised, and the company’s own boilerplate now lists only Singapore and Australia as markets for its care services. “Global rollout” will need a map before it means much.

Still, the problem Esteem targets is real and growing. Southeast Asia’s next healthtech winners will be built around workflows, not just AI, and Japan, further along the ageing curve, is already turning to physical AI to keep eldercare running. If Esteem gives care managers back a few hours a week, families will feel the difference. If it mainly lets agencies do more with fewer people, carers will feel it too.

Esteem AI is available to care providers in Singapore from today.

The post Homage turns its in-house care software into a business as Singapore turns super-aged appeared first on e27.

Anyone in Singapore who has sat through a corporate health screening knows the ritual. A thick report arrives, a few numbers glow red, and the PDF is filed somewhere between a tax notice and a gym membership nobody uses. Add a smartwatch counting steps, a blood pressure monitor bought after a scare, and clinic records sitting in systems the patient never sees, and the average person ends up data-rich and insight-poor.

Genie Health, a Singapore startup co-founded by former PropertyGuru CEO Hari V. Krishnan, wants to close that gap. The company has raised US$3 million in seed funding in a round led by the Nadathur Family Office, with participation from Liminal, a Temasek-backed founders’ studio, and two unnamed healthcare leaders.

Also Read: Homage turns its in-house care software into a business as Singapore turns super-aged

Anand Nadathur, Principal of the Nadathur Family Office, joins the board.

The founding team is unusually heavy-hitting for a seed round. Krishnan led PropertyGuru’s NYSE listing and its subsequent US$1.1 billion sale to EQT, and before that ran LinkedIn’s Asia Pacific business. Co-founder and Chair Maya Hari was CEO of carbon accounting firm Terrascope and earlier led Twitter’s business in the region. Sandeep Suvarna headed international marketing at Twitch and Asia Pacific marketing at LinkedIn.

Investors are plainly backing the people as much as the product. Nadathur said his office backs founders who are “capable, committed, and aligned with our values for the long term.” What the trio does not have, at least among its co-founders, is a clinician. The two healthcare angels and the planned hiring of healthcare expertise look designed to fill that gap.

Two products, one bet

Genie Health pitches itself as a two-sided health intelligence platform. On the consumer side, the free Genie app pulls health screening reports, lab results, wearable data, connected devices and family health history into a secure “Health Locker”, then surfaces personalised insights and suggests next steps. Members, the company says, keep control of their data at all times.

On the provider side, Genie SmartCare is a clinic platform built on agentic AI, software that carries out multi-step tasks rather than simply answering questions. It engages patients around the clock, books appointments and follow-ups, and combines longitudinal health records with wearable data so care teams see a fuller picture. Providers, the company stresses, retain complete oversight.

The theory is that each side feeds the other: consumers bring data, clinics bring trust and care pathways. The new money will deepen the platform’s AI, push SmartCare into more clinics and grow product, engineering and healthcare teams in Singapore.

Also Read: Why Southeast Asia’s next healthtech winners will be built around healthcare workflows, not just AI

The demand case is real. According to the company, one in three Singapore residents lives with at least one chronic condition, such as high blood pressure, high cholesterol or poor glucose control. As the population ages, primary-care teams are carrying heavier loads, and the government’s Healthier SG programme has pushed preventive care to the centre of national health policy. Investors have taken note: listed players are already testing investor appetite for Singapore’s ageing economy.

“Members tell us they have more health information than ever, but still struggle to understand what it means for them and what to do next,” Krishnan said.

The hard parts

Aggregating health data sounds simple and has humbled many who tried. The app is rarely the difficult bit; making records portable across hospitals, clinics and labs is. Much of Singapore’s clinical data sits within public healthcare clusters and national systems, so a consumer app’s insights are only as good as what users upload and what partners agree to share. As one e27 contributor argued, the future of healthcare AI isn’t more data but better context.

Then there is trust. A locker holding lab results and family history is a prized target, and Singapore remembers the 2018 SingHealth breach, which exposed the records of about 1.5 million patients. “Members retain control” is a promise that will need audits, certifications and transparent data practices behind it.

Personalised guidance also sits close to medical advice. The line between spotting a pattern in someone’s cholesterol readings and telling them what it means clinically is thin, and the Health Sciences Authority, which regulates software as a medical device, will care where Genie draws it.

On the clinic side, SmartCare’s agentic promise meets the messy reality of primary care, where bookings still live in phone calls, WhatsApp threads and front desks. Patient intake has become one of healthcare’s most important AI use cases precisely because it is so manual, but AI agents in healthcare only work when the hand-off to humans is designed carefully.

“Healthcare teams want more time to understand their patients and support them prior to and beyond the consultation, but routine administrative demands can get in the way,” Maya Hari said.

A crowded waiting room

Genie Health enters a busy field. The government’s HealthHub app already gives Singapore residents free access to public health records, appointments and Healthier SG enrolment. Telehealth players such as Doctor Anywhere and MaNaDr offer consultations alongside digital records.

On the clinic side, Zeya Health recently acquired ConnectLah to build AI booking rails, and hospitals in the region, such as this Bangkok hospital, are building agentic patient services in-house. Globally, Apple Health and Google’s Fitbit already aggregate wearable data, while US startups such as Function Health have raised large sums selling lab-driven personalised insights.

Genie’s edge, if it has one, is the two-sided design and its founders’ platform DNA. PropertyGuru, after all, was a marketplace that kept the portal free for home seekers and earned its money from agents. Genie’s consumer app is also free, which suggests clinics will foot the bill, though the company has not disclosed its business model or how many clinics are using SmartCare.

Also Read: Japan’s I.W.G raises US$1.8M to stitch Asia’s fractured medical records

That playbook could travel well. But health is not property. A misleading listing costs a buyer a wasted viewing; a misleading health nudge costs a great deal more. Krishnan’s team has proven it can build a platform people use. Its next test is building one people trust with their bodies.

The post Ex-PropertyGuru CEO’s Genie Health raises US$3M to make sense of your scattered health data appeared first on e27.

For years, the small business owner in Penang or Cebu has relied on a quiet, unspoken form of security: obscurity. Who would bother hacking a family-run online furniture store when there are banks to rob?

The logic was not wrong. Breaking in took skill, time and patience, and skilled criminals price their time like consultants. They went where the money was.

That logic died this month, and it cost about US$50 to kill it.

Also Read: Singapore Polytechnic launches CASTLE to help SMEs strengthen cyber defences

According to a report by NorForge researcher Snorre Fagerland, first reported by Reuters on 19 October, a single Chinese-speaking operator used an open-source agent platform called Cairn to break into the websites of more than 100 companies, including Deere, SharkNinja and Cummins. The agents attempted more than 1,200 sites.

The operator’s instructions ran to fewer than 100 Chinese characters. The agents did the rest, taking a little under 23 hours per successful intrusion. Fagerland estimated the whole campaign burned roughly US$5,000 in tokens, or about US$50 per break-in.

“It’s not script kiddie stuff,” Fagerland told Reuters. Quite. It is something worse: script kiddie economics with professional-grade output.

Not the giants. The middle

The headlines named the American giants because giants make headlines. The more instructive detail sits lower down. NorForge found that the busiest cluster’s typical target was an online shop with traffic of between 200,000 and 1.5 million, in the US, Canada, Japan and Australia. The goal was to plant a web shell for lasting access to shop and payment systems, to be cashed in later.

Read that profile again with Southeast Asian eyes. It describes the region’s mid-sized direct-to-consumer brands: the beauty labels, gadget resellers and fashion startups that grew up on marketplace traffic and are now building their own storefronts.

Southeast Asia wasn’t on this operator’s list. That is luck, not design. Nothing about the method is tied to geography, and an operator already working through Japanese checkout pages will not find Bahasa Indonesia or Thai ones any harder.

The structural problem is that the region’s digital economy grew faster than its defences. An ERIA study cited by Positive Technologies found that only 68.5 per cent of small businesses in ASEAN use specialised software to protect their infrastructure. Roughly three in ten have nothing beyond whatever their hosting provider bundles for free.

The plugin you updated last Thursday

If that sounds abstract, consider the other security story of the week. Between 15 and 16 October, the distribution process for Rank Math PRO, a popular WordPress SEO plugin, was compromised. Some copies of version 3.0.123 carried code capable of creating persistent access to the sites that installed them. Administrators were told to treat their websites as potentially breached and restore from clean backups.

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

This is how small businesses actually get hurt. Not through elaborate zero-days, but through the chain of plugins, themes, freelance developers and forgotten admin panels that holds a typical SME website together.

The founder does not know which version of which plugin is live. The agency that built the site moved on two years ago. The admin password lives in a WhatsApp group.

AI agents are tireless at exactly this kind of work. A human attacker gets bored scanning 1,200 storefronts for a misconfigured control panel. An agent does not get bored. It does not ask for a salary either.

Enterprises get guardians, SMEs get a blog post

The industry’s response so far tells you who matters. This week Microsoft unveiled Guardian Agents, which supervise enterprise AI agents in real time and treat them as potential insider threats, with task-scoped permissions and human approval for high-impact actions.

It is sensible engineering. It also sits inside Microsoft 365 E7, which is not a product line the kuih supplier in Shah Alam is likely to be shopping in.

Meanwhile, the bill for getting it wrong keeps climbing. IBM’s 2026 Cost of a Data Breach Report put Southeast Asia’s average breach cost at a record US$4.12 million, 12 per cent higher than last year, and IBM partly blamed AI for making attacks faster and cheaper to run.

That figure comes from just 26 organisations, so treat it as a direction rather than a price tag. But the direction is unmistakable. And a small business does not need a US$4 million bill to be finished. A week of hijacked checkout pages and a few hundred leaked card numbers will do the job.

So the region is drifting into a two-tier security economy. Large companies will buy AI to fight AI. Everyone else will rely on hope and a free SSL certificate. When attack costs fall to US$50 and defence stays priced for enterprises, the gap is not technical. It is a market failure.

What a sane response looks like

Market failures are what platforms and governments are for, and three moves would matter more than another cyber-hygiene awareness campaign.

Also Read: From KYC to KYA: how AI agents are reshaping payment risk

First, push security down the stack. The marketplaces, website builders, payment gateways and cloud resellers that host Southeast Asia’s small sellers are the only players able to defend millions of merchants at once. Automatic plugin patching, default multi-factor login and alerts for exposed admin panels should be the baseline, not a premium tier. Payment providers have the clearest incentive of all: the Cairn operator was after payment systems.

Second, treat agent infrastructure as a loaded weapon left on the table. NorForge saw this campaign only because operators had left their Cairn control panels open on the internet, and those panels leaked hundreds of credentials and API keys belonging to victims. Hosting providers in the region should scan for and take down malicious orchestration panels as routinely as they remove phishing pages.

Third, hold the AI labs to account for where their tools end up. NorForge’s data showed tasks running on a worker labelled after Anthropic’s Claude Code and DeepSeek’s low-cost V4.1 Flash model, though the firm offered that as a reading of the label, not a confirmed stack. Whatever the precise combination, model providers selling into ASEAN should publish regional abuse data and fund defensive tools for the small companies their products are now being turned against. Regulators signing AI safety agreements with these labs should be asking for exactly that.

Obscurity was never a strategy

Southeast Asian founders have long been told that cybersecurity is a cost centre to sort out after product-market fit. That advice assumed attackers were scarce and expensive. They are now neither.

The uncomfortable lesson of the Cairn campaign is not that one hacker got into Deere. It is that one hacker, armed with a few thousand dollars of tokens and a sentence of instructions, could plausibly work through every mid-sized online shop in Jakarta while asleep.

Also Read: The new ransomware playbook: Why ASEAN banks are losing the disclosure war

The region’s small businesses were never safe because they were hidden. They were safe because they were not worth the effort.

At US$50 a door, everyone is worth the effort.

The post AI agents broke into 100 firms for pocket change. SEA’s online shops fit the target profile appeared first on e27.

In February 2026, Sea Limited struck a partnership with Google on “agentic shopping prototypes.” By June, Shopee had become the first Southeast Asian e-commerce platform built directly into ChatGPT, live across Indonesia, Malaysia, the Philippines, Singapore, Thailand, Taiwan and Vietnam. Lazada, meanwhile, built its own AI shopping agent, Lazzie, running on Alibaba’s Qwen model, and credits it with a measurable lift in engagement during flagship sales events.

Two of the region’s biggest commerce platforms have quietly redefined who, or what, the customer actually is.

The customer might not be a person anymore

For two decades, selling online in Southeast Asia meant winning a person’s attention: a good photo, a fair price, a five-star review a shopper could read for themselves. Increasingly, it means winning an algorithm’s judgment instead. When a shopper asks ChatGPT to find “a good baby stroller under SG$150 (US$117),” it is the agent that reads the listings, compares the options and applies whatever criteria it was built with. Only the winner it picks ever reaches the human.

Consumers seem ready for exactly this handoff, in precisely the categories Southeast Asia’s small sellers depend on most. A June 2026 Checkout.com study of consumer and merchant readiness found shoppers are most comfortable delegating grocery shopping (41 per cent) and household supplies (31 per cent) to an AI agent, the bread-and-butter categories for millions of small Shopee and Lazada sellers, while trust drops sharply for higher-stakes categories like financial services (15 per cent). The same study found 89 per cent of merchants are already preparing for agentic commerce, even though only 3 per cent of transactions today actually involve an AI agent. Readiness is running well ahead of reality, which is usually when the ground shifts fastest.

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

Selling to an algorithm nobody can audit

Here is the uncomfortable part for the region’s small sellers: nobody outside Shopee, Lazada, OpenAI and Alibaba can see how these agents actually decide what to surface. A merchant can optimise a listing for a human eye, with better photos, clearer copy, faster replies, because a human’s attention is at least partly legible. An AI agent’s ranking logic is not published, is not stable, and is not something a small seller has any real way to contest if it quietly stops recommending their shop.

The market is not waiting for sellers to catch up. The global AI shopping assistant category is projected to grow from roughly US$4.2 billion in 2025 to US$22.1 billion by 2032, and in Southeast Asia, AI recommendations are already estimated to influence 90 per cent of purchase decisions, with 80 per cent of shoppers engaging these tools weekly. That is not a future scenario. It is the market Southeast Asia’s sellers are already operating inside.

To be fair, this is not a story of platforms acting in bad faith. Shopee’s OpenAI partnership includes ChatGPT for Business trials for eligible sellers, meant to help them write listings, marketing copy and automate customer service. That help matters, but it addresses the seller’s side of the conversation, not the agent’s decision logic on the other end of it.

Even consumers, who stand to gain the most convenience, are not handing over full control: Checkout.com’s respondents wanted spending caps, instant revocation and easy cancellation before they would trust an agent to buy on their behalf, and on average would only let one spend around £177 (US$237) without asking first. If shoppers are this cautious about what an agent can do with their own money, sellers are entitled to at least the same caution about what an agent can do to their livelihood.

Also Read: The hidden economics of autonomous AI agents

The capital isn’t going where the sellers need it

It would help if the region’s AI investment boom were flowing toward the unglamorous work of making these systems legible to the small businesses that now depend on them. Mostly, it isn’t. Southeast Asia’s native AI startups raised a record US$4.1 billion in the first seven months of 2026, more than double all of 2025, but roughly 68 per cent of that sits in a single transaction, Kling AI’s US$2.8 billion round for its generative-video platform. Strip that one deal out and regional AI funding is closer to US$1.3 billion, concentrated in infrastructure and foundation models, not in tools that would let a small seller understand why an agent stopped recommending their stall.

I did not train as an engineer. I trained as a paramedic, then taught myself to build software because I wanted ordinary people, not only institutions, to have a fair shot at seeing what’s coming before it arrives, an instinct that eventually carried me from self-taught builder to founder and, later, to a board and product-architecture role at a Singapore culture-tech venture. That background is why this particular shift concerns me more than most AI headlines do. The risk was never that AI agents would simply replace shoppers. It’s that they already have, in exactly the categories where the region’s smallest sellers make their living, and almost nobody has given those sellers a way to see, question, or appeal what the agent decided.

Southeast Asia spent a decade teaching hundreds of thousands of small sellers how to win over a customer. It now owes them a fair shot at winning over the algorithm that has quietly taken that customer’s place.

—

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 Southeast Asia’s sellers spent a decade learning to sell to people. Now they have to sell to AI appeared first on e27.