Most forecasts of enterprise AI in Southeast Asia picture it arriving the way ERP did: a board decision, a transformation budget, a systems integrator, a two-year rollout. We think that picture is wrong for this region. Over the next three years AI will come into Southeast Asian business through the side door, one professional at a time, and the companies built for that buyer are the ones worth backing now.
Start with who the buyer actually is
Well over nine in ten businesses across ASEAN are micro, small or medium enterprises, and together they employ most of the region's workforce. The typical business here is not a listed conglomerate with a CIO. It is a clinic, an accounting practice, a freight forwarder, a distributor, a family manufacturer with a hundred staff.
Those businesses do not have a transformation budget or a procurement committee. They have an owner who signs off on spend, a handful of professionals who carry the real workload, and a pile of documentation, reconciliation and reporting that gets finished after the working day is supposedly over.
That is the buyer. And that buyer does not run a request for proposal. They try something on a Tuesday because a peer mentioned it, and they keep paying for it if it gave them an evening back.
Why the top-down playbook stalls here
The enterprise playbook assumes a sponsor high in the organisation, a budget line, and an IT function that can integrate a new system into an old one. In most Southeast Asian companies at least one of those three is missing.
The large enterprises that do have all three are real buyers, and some of them will run serious programmes. But they are fewer, their cycles are long, and in several markets the procurement process favours incumbents and local integrators over a young software company. A seed-stage founder who builds only for that buyer is signing up for long sales cycles on a runway that early-stage capital in this region is not generous enough to extend.
The side door is faster. The person doing the work feels the pain every day, has authority over their own tools, and can see within a week whether something helps.
What the side door rewards
Winning a professional who buys for themselves is a different design problem from winning a committee. Three traits keep showing up in the products we think will carry this wave.
First, they work inside what is already there. A professional will not migrate their records, change their templates or retrain their staff to try a tool. The product has to sit alongside the existing system, whatever it is, and ask for nothing on day one. If adoption needs a change-management plan, the side door is closed.
Second, the output can be checked. A doctor signs the note. An accountant signs the filing. A professional will only hand work to software if they can see what it produced, correct it quickly, and put their name to it. Output that cannot be checked gets used once and then quietly worked around.
Third, they handle how people here actually work. Conversations in Manila, Kuala Lumpur and Jakarta regularly move between two or three languages in the same sentence, and documents follow local formats set by local regulators. Tools trained and tested somewhere else often break on exactly that texture, and the professional notices on the first day.
Built regional from the start
This is where the side door connects to something we believe about building in this region more generally. Southeast Asian B2B companies do not usually stall by going regional too early. They stall by waiting too long, winning a home market that turns out to be too small to carry the company, and then discovering that expansion is a build they never started.
A product designed for mixed-language, locally regulated professional work is already most of the way to its second market. The underlying job of a clinic in Cebu and a clinic in Selangor is close enough that a team solving it well in one should be planning for the other from the first year, with the local formats, the language handling and the regulatory detail treated as part of the product rather than an afterthought.
That is also where a seed investor can be useful beyond the cheque. Cross-border expansion support, the introductions and the practical detail of entering a second and third market, is a large part of what we do for the companies we back.
What happens to the big transformation budgets
They do not disappear. We expect large enterprises in the region to keep spending on AI, and some of that spending will be well directed. But we expect a good share of it to follow the side door rather than lead it.
When enough professionals inside a hospital group or a mid-sized bank are already using a tool on their own, the enterprise deal stops being a leap of faith and becomes a formalisation of something that already works. The company that earned those individual users first negotiates that deal from a very different position from the one cold-calling the CIO.
What this means for how we underwrite
When we look at an AI company selling into Southeast Asian business, we ask four questions before we ask about the model.
Who is the first person to pay, and do they have the authority to pay without a committee? What expensive, manual work does the product take off their desk, and can they tell you unprompted how much they resent doing it today? Can they check and sign what the product produces? And is the team building for a second market already, or treating it as a reward for winning the first?
A company with clear answers to those is building for the buyer this region actually has.
The challenge
If you are building AI for Southeast Asian business, write down the name and job title of the first person who will pay you, and the evening they get back. If the answer is a department rather than a person, it is worth asking whether you are waiting at the front door of a building where most people come in through the side.