Main idea

Connecting tools, indexing documents, generating notes, building a graph, plugging in an assistant: each of those operations is now a solved problem, with vendors, budgets and known lead times. That is not where projects fail.

What blocks comes down to concessions nobody wants to make out loud. Accepting that the delivered document is not what deserves to be kept. Accepting that raw information has to be reworked before it is worth anything, hence that collecting it is not enough. Accepting that a company's memory cannot stay in the heads of a few people — which touches directly on those to whom that memory gave their standing. Accepting that context is built, tended and governed, hence that it costs time without producing a visible output.

The blockage is a matter of posture and status, not of tooling. And it is not lifted by a demonstration: no technical proof answers a reluctance that is never stated.

Why it matters

It explains why a technically flawless capitalization project can never be adopted, and why the failure will be analysed as a tool problem — the only cause you can talk about in a meeting.

It also says where to put the effort at launch: on what the organization agrees to recognize as work, before the platform.

Nuances and limits

The claim only holds once the technical means are genuinely available. Where information is locked inside tools with no programmable access, the obstacle really is technical, and culture can do nothing about it.

And "cultural" does not mean irrational: refusing to share knowledge that grounds your own value in the organization is a sound calculation, as long as nothing rewards sharing.

Open questions

• What form of work recognition would make tending a shared memory stop being a bad individual calculation?