Skip to content
the ai reality map · course 01 · chapter 02/10
// 02 · the use-case

Where does AI actually matter, and where is it theatre?

Short answer: buy the commodity, build only where you own the data and the workflow, and wait if the data is not ready. That one rule filters out most bad AI spend, and the tool below gives you the verdict in a minute.

Build, buy, or wait is the real decision. MIT found bought or partnered AI reaches production about 67 percent of the time versus 33 percent for internal builds (MIT Project NANDA, 2025).

Pick the place where you own the data and the workflow. Everything else is someone else's commodity.

// build · buy · wait

The real first decision is not which model. It is whether to build, buy, or wait. Answer honestly and the tool gives you the same verdict we would.

Do you own data or a workflow a competitor could not copy?
Is that data clean, accessible, and legal to use today?
Does an off-the-shelf tool already do most of this well?
Can you name the metric it must move, and where it sits today?
Your company size?

// the short version

share
transcript
  • The real AI decision is build, buy, or wait.
  • Bought or partnered AI reaches production about twice as often as internal builds (MIT).
  • Build only where you own the data and the workflow.
  • Run the free, interactive course at heimlandr.io/ai-reality-map.

// the deep dive

Buy the commodity (transcription, generic chat, coding assistants); building it yourself is lighting money on fire. Build only where you have proprietary data or a workflow nobody else can copy, because that is the only place a custom system earns its keep. MIT's 2025 data is blunt on this: tools bought from specialist vendors or built through partnerships reached deployment about twice as often as tools built internally. And sometimes the right answer is wait: if your data is not ready, a pilot now just produces an expensive lesson you could have read here for free. Use the tool above to get the same verdict we would give you.

// chapter faq

When should a company build custom AI instead of buying?

Only where it owns proprietary data or a workflow nobody else can copy, because that is the only place a custom system can become an advantage. Everywhere else, a bought tool ships faster and fails less: MIT found bought or partnered AI reaches production about twice as often as internal builds.

Which AI should every company just buy off the shelf?

Transcription, generic chat and coding assistants. They are commodities: excellent, cheap, and identical for you and your competitors. Building your own version of a commodity is lighting money on fire.

When is waiting the right AI decision?

When the data is not clean, accessible and legal to use yet. A pilot on unready data just produces an expensive lesson; spend the quarter fixing the data instead, then build from strength.

Every figure in this chapter is sourced. The full source list lives on the main map. Open the map

This is one chapter of ten. The whole course is free.

The full map has the interactive tools, the 8 minute audio edition, the live layer and every source. And if you want it run against your own reality, that call is free too.

Open the whole map