Ten AI Startups I Would Build If I Were 25

People collect AI startup ideas like trading cards. Ideas are cheap. What is scarce is a wedge, a specific reason a small team can win against incumbents and against general models. So here are ten I would build if I were twenty-five again, each chosen for a real wedge, and biased toward Bharat, where the problems are large and the tools are finally good enough.

Ten I would build

1. Vernacular voice agents for the frontline. Millions of workers and customers speak rather than type, in a dozen languages. Voice-first agents for field sales, support and public services, in Indian languages, are a wedge global players will not prioritize.
2. Compliance back office for regulated SMEs. GST, filings, audits: painful, rule-bound, high volume. A vertical agent that reads, checks and files is dull and durable.
3. Agentic leverage for solo legal and accounting practices. Not to replace them, to give a two-person firm the reach of twenty.
4. Diagnostic vision for under-served clinics. Screening from images where specialists are scarce, built with the humility and traceability healthcare demands.
5. An AI layer over informal credit. Underwriting the thin-file borrower from alternative, consented data. The data moat is real and local.
6. Agri intelligence per plot, not per region. Sowing, disease and market timing specific to a farmer's field, delivered by voice.
7. A workflow-owning agent for one boring vertical (freight, claims, logistics documentation). Sell the outcome, not the model.
8. On-device AI for privacy-sensitive work. Where data cannot leave the building, a small-model product wins by default.
9. An eval and reliability platform for AI teams. When everyone builds agents, the picks-and-shovels play is helping them trust the agents.
10. Talent infrastructure for the AI era. Mapping real skills to real work, and closing the gap, which happens to be what I am building with GuildTrek.

The pattern behind the list

Notice what these share. Each owns a workflow, not a model. Each sits on proprietary or local data a general model cannot reach. Each solves a painful, high-volume problem for a specific user. None of them is "a better chatbot." That is the wedge. Ideas are free. Wedges are earned.

So of these ten, which one is sitting in your industry, waiting for someone who actually understands the domain to build it?