We spent March through the end of May in the Block71 x Microsoft AI Accelerate program, a 10-week cohort run by Block71, NUS Enterprise, and Microsoft, with support from Enterprise Singapore through the Digital Industry Singapore (DISG) office. Our track was called "Scaling AI Solution in Asia," and it was aimed at founders building AI products toward real enterprise revenue in the region. Here's an honest write-up of what it was and what we got out of it.
What the program is
It's part of a bigger commitment from Microsoft, Enterprise Singapore, and NUS Enterprise to put up to 150 AI startups through this over the next three years, with a smoother path into the Startup SG Tech grant. The cohort kicked off with three days in early March and then ran for ten weeks of masterclasses, 1:1 clinics, and mentor sessions, ending with an Investor Day in late May.
A few things stood out as genuinely useful. The masterclasses covered AI product strategy and enterprise go-to-market, FinOps, getting production-ready on Microsoft Foundry, plus the less glamorous but necessary stuff like foreign incorporation, legal structuring, and fundraising. On the distribution side there's Microsoft co-sell and marketplace eligibility, and a clear route into Enterprise Singapore's grant tracks (up to S$400k for a proof-of-concept and up to S$800k for a proof-of-value).
The three-day kickoff
The opening three days (March 3 to 5) ran roughly in three arcs:
- Day 1, foundations. The Microsoft AI and agentic landscape, the case for Singapore as a regional HQ, goal-setting for the sprint, and getting paired with mentors.
- Day 2, building the engine. FinOps with Airwallex, AI product strategy and enterprise GTM with Summer Yang, and a fireside with founders who were already further down the road.
- Day 3, going global. Production readiness and Microsoft Foundry, foreign incorporation with Osome, and legal structuring and fundraising with Gunderson Dettmer.
One line from Summer Yang at Microsoft ASEAN has stayed with us: "Building fast is no longer the hard part. Scaling right is." Her point was that with AI tooling, shipping a product is the easy bit now. The hard part is baking go-to-market, trust, and distribution into the product from the start instead of bolting them on later.
What we actually took away
Three honest things, in case you're weighing up a program like this:
- You get out what you put in. The structure is solid, but how much you get back depends almost entirely on how hard you push it.
- The people were the real win. A few of the other founders in the cohort are now GeoVector users, and the conversations with them ended up more valuable than any single session.
- It accelerates what's already moving. A program like this multiplies momentum you already have. It won't manufacture it from scratch, which is worth knowing going in.
Thanks
Thanks to Block71, NUS Enterprise, and Microsoft for having us, and to the other founders in the room for making it worth the ten weeks.
Curious what AI assistants are telling buyers about your business, and where your competitors are winning the shortlist? Get in touch with our team.