It was a Sunday afternoon. A founder I know, running a company with 100 employees, revenue above US$50 million, was mid-pickleball when a message arrived: “Can you speak at our conference Tuesday morning?”

He agreed. The conference was two flights away.

What followed was two and a half hours across six browser tabs. Flight aggregators. An airline app. Hotel comparison sites. His loyalty portal. Google Maps for the airport transfer. A WhatsApp message to a friend asking which hotel was actually good near the venue.

By the time he’d confirmed everything, he’d made fourteen booking decisions, compared forty-seven options, and completed a job that neither his role nor his Sunday afternoon were designed for.

That’s not a story about bad travel tools. It’s a story about a category that was designed, from its inception, to serve someone else.

We assumed the problem was price, we were wrong

When we started Bliink, the working thesis was simple: the economics of corporate travel management were broken. Legacy TMCs like Amex GBT, CWT, and BCD require minimum annual travel spend of US$500,000 to access managed services. That threshold eliminates virtually every SME in Southeast Asia before the conversation begins.

So we built for the SME tier. And then we started listening.

After working directly with companies across Indonesia and Singapore, the answer that kept surfacing wasn’t price. It was time. Convenience. And fragmentation.

Companies weren’t failing to book travel because they couldn’t afford a TMC. They were losing hours every week because no product had ever actually taken the job away from them.

Also Read: Why the US tech rebound matters for SEA’s AI and venture ecosystem

Three solutions, three failures

The corporate travel industry has had thirty years to solve this and has produced three categories of response. Each one fails differently.

• Legacy TMCs price you out.  Amex GBT runs an 80 per cent opex ratio, US$479 million in operating costs against US$597 million in revenue, managed by 19,000 people. That cost structure doesn’t support SME pricing. It never will. For Southeast Asia’s 60 million-plus SMEs, these platforms are structurally inaccessible.
• Self-serve tools still make you do the work.  TravelPerk, Navan, and SAP Concur gave companies a better interface and called it a solution. But a faster booking tool is still a booking tool. You still search, filter, compare, approve, reconcile, and report. The friction moved from phone to screen. The labour did not move at all.
• The SME market was never the target.  This is the one the industry has never said out loud: no travel management company was ever designed for companies with 20 to 500 employees booking in rupiah, using Lion Air, needing WhatsApp approvals, and managing travel across multiple time zones and currencies. Southeast Asia’s SMEs weren’t underserved. They were structurally excluded. There is a difference.

Why the window just opened

Three things have converged that have not converged before.

AI has made fully managed travel economically viable at SME scale for the first time. The 1:120 service ratio that legacy TMCs achieved by employing thousands of agents can now be delivered by a single AI system. What required a US$500,000 minimum spend in 2022 can be delivered at the 50-employee level in 2026.

Also Read: The future of marketing isn’t about AI, it’s about judgment

The legacy players are collapsing under their own weight. CWT filed for bankruptcy. Amex GBT is being taken private in a US$6.3 billion deal, which is a cost restructuring play, not a growth strategy. The managed travel market for SMEs is, for the first time, structurally undefended.

Southeast Asia’s SMEs have already digitised their behaviour. WhatsApp penetration exceeds 90 per cent in Indonesia. Mobile corporate expense adoption has tripled since 2022. The behavioural foundation that makes AI-native corporate travel workable exists in the market right now. It didn’t three years ago.

What the data reveals

We built Bliink to test this thesis. Eight months in, the metric that surprises people most isn’t revenue. It’s retention.

We’ve had 100 per cent client retention since the beta launch. We’ve never run a marketing campaign. Every client arrived through a referral.

I’m careful about what conclusions to draw from early data. But 100 per cent retention across eight months of operation is not a product metric. It’s a signal about what category you’re actually in.

There’s a difference between software that people use and a service that people keep. Corporate SaaS benchmarks 5-10 per cent monthly churn as normal. The gap isn’t in features. It’s whether the job belongs to the product or to the person.

When a platform handles the trip, preferences recalled, policy applied, itinerary sent, receipts filed, the work is no longer the traveller’s. That’s not a UX improvement. It’s a category shift. And it turns out that when you actually complete the job for someone, they stop looking for alternatives.

The actual opportunity

The opportunity in corporate travel intelligence for Southeast Asia is not better booking software. It’s institutional memory.

Every trip a company takes encodes information: traveller preferences, policy boundaries, pricing benchmarks, vendor performance, patterns across teams and time. That data doesn’t exist in any structured form for the 60 million-plus SMEs in this region. It’s scattered across email threads, WhatsApp chats, and booking confirmation PDFs that nobody reads twice.

The company that captures and structures that data, not for MNCs, not for Fortune 500 procurement teams, but for the mid-tier Indonesian consultancy and the Singapore regional distributor and the Jakarta family office, builds a moat that no generalist AI model or offshore OTA can replicate.

The Internet gave companies infinite choices and left them flying blind. AI is finally giving them back the trusted advisor the Internet took away.

The question for Southeast Asia is who builds it first, and whether they build it from here.

The post Corporate travel in Southeast Asia was never broken, it was never built appeared first on e27.