Unit 5: The build-buy-wait framework for AI capabilities

Learning objectives

Apply the build-buy-wait framework to AI investment decisions. Evaluate when proprietary AI features create lasting differentiation. Time AI investments against technology maturity curves.

Video script

Reading material

Decision criteria in detail

Differentiation test. Ask: if a competitor added this same AI capability tomorrow using off-the-shelf tools, would it erode our competitive position? If yes, you need to build something that off-the-shelf tools can't replicate (usually because of your proprietary data or unique integration). If no, buying is fine.

Data moat test. Ask: does this feature get better because of data that only we have? If your product generates unique behavioral data, transaction data, or domain-specific content that improves the AI's performance, building creates a compounding advantage. Each user makes the AI better, which attracts more users.

Technology trajectory test. Ask: will the technical cost of this capability decrease significantly in 6-12 months? Check the AI lab roadmaps, benchmark improvements, and cost trends. If the trajectory suggests rapid improvement, the wait option becomes more attractive.

Opportunity cost test. Ask: what else could our team build with the time and resources this AI feature would consume? If the alternative uses of the team's time are more valuable, buying or waiting makes sense even if building is technically feasible.

Portfolio approach to AI investment

Rather than making a single big bet, structure your AI investments as a portfolio across the three timing horizons.

60% in mature AI capabilities (act now). These are well-understood technologies where you're adding proven value to your product. Buy where possible, build where differentiation requires it. Ship in the current quarter.

30% in emerging capabilities (experiment and learn). These are technologies that are improving fast but aren't reliable enough for production commitments yet. Run experiments, build prototypes, test with users. Learn what works without committing to production-grade infrastructure.

10% in early-stage capabilities (watch and prepare). Stay informed about what's coming. Test new models and APIs when they release. Build organizational knowledge. But don't make roadmap commitments.

Practical exercise

Exercise: Build-buy-wait analysis

Choose three AI capabilities that a product you know could benefit from. For each one, run through the decision framework.

1. Is it core to differentiation? (Differentiation test)
2. Does it benefit from proprietary data? (Data moat test)
3. Will it be significantly cheaper/easier in 6 months? (Technology trajectory test)
4. What's the opportunity cost? (Opportunity cost test)

Recommend build, buy, or wait for each. Write a brief justification for each recommendation.