What's realistic by company size, enterprise to small?
Short answer: it depends on size, but not the way vendors say. Small firms should buy good tools; mid-sized firms earn one custom build on their real edge; enterprises need governance before scale. The divide is leaders versus laggards, not size.
Adoption is near-universal, impact is not: 88 percent use AI somewhere, only about 39 percent see any EBIT impact, and roughly 5 to 6 percent capture value at scale (McKinsey; BCG, 2025).
Enterprise, mid, SMB, and small should not do the same thing. Most should buy the commodity and build only where they're different.
›transcript
- Adoption is near universal. Impact is not.
- 88% use AI; only about 5% capture value at scale (McKinsey, BCG).
- Buy the commodity. Build only where you're different.
- Run the free, interactive course at heimlandr.io/ai-reality-map.
// the deep dive
A 20-person company that builds its own model is making an expensive mistake; the right move is to adopt good tools well and move on. A mid-sized firm earns a custom build only on the one workflow that is genuinely its edge. An enterprise needs the boring layer most vendors skip (governance, access control, a platform) before any single use-case is worth scaling. On timelines: Deloitte's 2025 survey found only 6 percent of organizations saw AI payback in under a year, with most reaching satisfactory ROI in two to four years, against the seven to twelve months typical of conventional tech. BCG found only 5 percent capturing value at scale and put about 70 percent of the gap on people, process and culture rather than the models. The real divide is leaders versus laggards, not company size.
// chapter faq
What should a small company do about AI?
Adopt good tools well and move on. A 20-person company that builds its own model is making an expensive mistake; the win at that size is disciplined adoption plus, at most, one narrow build on something genuinely unique.
What is a realistic AI roadmap for a mid-sized company?
Buy the commodity across the company, then earn one custom build on the single workflow that is genuinely your edge. Scope it to a metric you already track and expect payback in years, not quarters.
Why do enterprises struggle to scale AI?
Because the boring layer is missing: governance, access control, audit, a platform. BCG put about 70 percent of the gap on people, process and culture rather than the models. Without that layer, every pilot stays a pilot.
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.