Build, Buy, or Skip: A Framework for AI Decisions

Whether to build, buy, or skip an AI feature comes down to four questions, asked in order. Is this your actual differentiator? What does “good enough” cost to buy today? Does a free tier solve it for six months? Who owns maintenance once it ships? Build only when the capability touches proprietary data or a workflow no vendor matches. Otherwise, buy the commodity or skip it this quarter.

One pattern causes bad decisions here, every time. Teams score the build option on its upside: control, differentiation, no vendor lock-in. They score the buy option only on its downside: cost and lock-in. They also worry it won’t fit their workflow. Nobody scores skip at all, because nobody wants to be the person who said no to AI. A fair framework weighs all three the same way. The four questions below run that comparison.

Which four questions decide build vs. buy vs. skip?

• Is this your differentiator? Say the AI capability is core to why customers choose you: a competitor can’t easily copy it. Then building in-house is usually right, because you want to own the pace of improvement. Say it’s a capability every company in your category needs: document search, a support chatbot, spend visibility. Then you aren’t differentiating by building it yourself. You’re paying an ongoing tax for a slightly different version of something a vendor already sells.
• What does “good enough” cost to buy, right now? This is where teams skip the actual research. Before scoping a build, price the two or three closest off-the-shelf options at your actual usage volume. Skip the sticker price. Count seats, overages, and the integration work either option requires. We built CostMon specifically because this number is surprisingly hard to get straight. Spend is scattered across a dozen cloud, AI, and SaaS billing consoles, each with its own format. So “what would this cost us” ends up as a guess, not a number. Get the real number before you compare it to an engineering estimate that’s usually optimistic.
• Does a generous free tier already solve this for six months? Before you build or sign a contract, check whether a free tier is enough. It should cover the next two quarters at zero cost, long enough to learn whether the need is real. FreeTier exists because most teams don’t memorize the free tiers of 1,100+ tools. Teams should check what’s free first, but skip that step when they’re excited about a new build. Skip isn’t a permanent no. It’s often the right answer for this quarter. Revisit once you have real usage data instead of a guess.
• What’s the actual maintenance burden once it’s live? A build’s cost doesn’t end at ship date. Every AI feature needs monitoring, periodic re-evaluation as models and usage patterns change, and someone accountable when it degrades. Vendors amortize that cost across every customer they have; a bespoke build puts it entirely on your team, indefinitely. If nobody can name who owns that maintenance six months from now, that’s a real cost that belongs in the build-vs-buy math, not an afterthought.

When is building it yourself the right call?

Build wins in three cases. The capability touches your proprietary data or a workflow no vendor’s product matches. The volume is high enough that a per-seat or per-call vendor price costs more than engineering time. The integration is deep enough that a bolt-on tool needs almost as much custom work as building it yourself. Take a modeled illustration. Once a per-seat vendor tool crosses roughly 200–300 seats at typical AI-SaaS pricing, its annual license cost can exceed a senior engineer’s loaded time. That time would cover building and maintaining the same feature. It’s a useful gut check, not a universal rule. In those cases, build it. Give it the same production discipline as anything else you ship, not as a side project.

When should you buy or skip instead?

Buy wins for capabilities that are commoditized. Think model access, vector searchVector search finds the nearest matches to a query by comparing embeddings, not exact keywords. infrastructure, and observabilityObservability captures traces of every prompt, retrieval, tool call, and response for debugging. tooling: every AI feature needs them, but none of them is the product. Skip wins more often than founders expect. The use case might be speculative. The team might not yet have data on what “good” looks like. Or a free tier or a 30-day trial might answer the question more cheaply than either alternative.

What is a build-vs-buy-vs-skip review for?

The goal of a build-vs-buy-vs-skip review isn’t to arrive at “build,” “buy,” or “skip” as an ideology. It’s to make the tradeoff explicit and defensible, with real numbers instead of a preference dressed up as a strategy. That’s the shape of our strategy and architecture engagements. They typically run two to three weeks of diligence. The result is an honest read on what’s worth building, what’s worth buying, and what’s fine to leave alone this year. We back it with a written recommendation your team and board can act on.