Why Not Everyone
The idea is not the opening — the ugly, liability-bearing execution is, and broad platforms are structurally bad at it.
Objection. If the commercial memory layer is the obvious thing to build, why isn’t everyone already building it? And: who is attempting it, so we can study their screens, UX, and what users dislike?
Answer, half 1: nothing stops them. “Nothing, in theory. Assume they will.” Procore can build it. Document Crunch (now Trimble) is close to the commercial-risk side. Trunk Tools and Pype sit in the document-and-action layer. This is not secret whitespace — “it is the direction the market is moving.”
Spine line: “The idea is obvious. The execution is ugly.” The opening is five grinding execution problems big platforms are poorly shaped to solve:
| # | Reason | Why platforms are bad at it |
|---|---|---|
| 1 | Source data is fragmented and half-rotten — across Procore, WhatsApp, email, Excel, PDFs, photos, delivery notes, drawings, programmes, RFIs, variation logs, memory. | Generic AI search over documents “is not the same as building a normalized event history.” Cleaning and joining the mess is real, slow work — not magic. |
| 2 | Field adoption is brutal. A site manager “will only use something that is faster than their current habit.” | The bar is “Speak for 90 seconds. Add photos. Done.” Software head office loves dies on site. |
| 3 | Commercial judgement carries liability. A notice, claim event, or variation implication is not low-risk. | Forces human approval, source citations, explicit uncertainty → rules “no source, no claim” and “human approval before contractual output.” Platforms optimising for volume avoid this friction. |
| 4 | Historical cost memory is messy — final accounts, cost codes, subcontract orders, scope, inclusions/exclusions are inconsistent. | ”Build this too early and you produce confident nonsense.” |
| 5 | Big platforms optimise for broad workflows. Procore, Autodesk, Trimble need workflows for everyone. | ”They may not go deep on one specific mid-market commercial contractor workflow unless there is enough demand. That is your opening.” |
Reframe. The bet is not “we thought of something nobody did.” It is: do the ugly, narrow, liability-bearing, adoption-hard work a broad platform is poorly shaped to do, for a buyer the platforms under-serve. Clever ideas get copied in a sprint; ugly execution moats take years.
The objection also set the research agenda: study the competitors — screens, pricing, what users love and hate. That program is what the next chapter tests, the thesis.