In Manila and Jakarta, a parent searching for a digital loan or a small business owner hunting for a mobile wallet may never see the fintech company that could help them. The reason sits in the raw code of the company’s homepage — specifically, what isn’t there until a browser runs JavaScript. A new study by conversion specialist Slobodan Manic measured 274 fintech homepages from the CNBC World’s Top Fintech Companies 2025 list.
He took two readings on each site: a raw HTTP fetch with no JavaScript execution, and a full browser render. The gap between them is the gap an AI agent has to close on its own.
For 36 percent of those websites, that gap is too wide. Their homepage delivers less than 80 percent of its content in raw HTML. This matters because AI agents increasingly power search and information retrieval.
They do not automatically run JavaScript. They get the raw version first and decide whether to spend compute on a full render.
Most do not. If the page looks empty, the agent moves on. The core issue Manic identifies is something he calls “rendering independence.” Critical information, he argues, must not depend on client-side JavaScript.
Until last week this was a design principle. Now it is a number, and the number is uncomfortable. For Southeast Asia, the stakes are concrete.
Fintech has become a lifeline for unbanked and underbanked populations across the region. Digital wallets, microloans, and payment apps have spread rapidly, often leapfrogging traditional banking infrastructure.
But if the very companies offering these services are not optimized for AI agents, they risk being invisible in the search results people increasingly use. Think of how a family in a provincial capital searches for a loan. They type a query into a search engine or ask a voice assistant.
That query triggers an AI agent. The agent fetches the raw HTML of a fintech company’s homepage.
If the page is mostly empty code, the agent treats it as low-quality or irrelevant. The company disappears from results. The family never sees it.
Manic’s research measured the top fintech companies globally. These are not obscure startups. They are the companies on the CNBC list.
A third of them fail this basic test. The problem is not obscure technical edge cases.
It is the most important page on the property — the homepage. The report was published on May 25. It is a snapshot of a single moment.
But the trend it captures has been building for years. As AI agents become the default interface for online discovery, the rules of visibility are shifting.
Search engine optimization once meant keywords and backlinks. Now it may mean raw HTML weight. Manic’s work frames this as part of something he calls Machine-First Architecture.
The Structure pillar of that framework states that critical information must not depend on client-side JavaScript. This has been a design principle for a while. Now it has a measurement.
The measurement says a third of the sector is not ready. For the parent in Manila or the business owner in Jakarta, the consequence is simple.
They may search for a financial service that exists, that could help them, and never find it. The AI agent skips it. The company stays invisible.
The gap between the service and the person who needs it is not a lack of infrastructure or regulation. It is a few lines of code.
