consulting / custom software

a Palantir-style audit/plan/make software business — you go into a company, audit what software they have and what they actually need, propose a plan, then build it. the model works because most companies are underserved by off-the-shelf software for their specific workflows, and the cost to build custom tooling has dropped dramatically with AI-assisted coding. what used to require a 6-month engagement can now be scoped, built, and deployed in weeks.

the interesting business model tension is between the consulting side (high-margin, relationship-driven, hard to scale) and the productization side (lower-margin per client but recurring revenue, scalable). the Palantir reference is deliberate — they built a specific workflow tool for every client, then productized the patterns that emerged across clients. the play here might be: start with a handful of clients in a specific vertical (e.g. small biotech firms, local schools) to build deep workflow knowledge, then productize the patterns into a SaaS product. the overnight app grinder and spec-driven dev kit are directly relevant infrastructure — the faster and more reliably you can build software, the more profitable the per-client engagement.