AI disclosure
Where AI is used
EUC Collective uses one AI system, in one place: an intake assistant, built on Anthropic's Claude, that members talk to when raising a problem. It is not used anywhere else on this site — not to write library content, not to answer support questions, and not to generate the recommendations that end up in the library.
What the intake assistant does
- Acknowledges the problem you've raised
- Asks clarifying questions to capture what a human would actually need — your environment, scale, constraints, what you've already tried, and what outcome you're after
- Summarises the problem back to you to confirm it's understood correctly
- Confirms it's been filed and tells you what happens next
What it deliberately does not do
- It does not answer your technical question
- It does not recommend a product, approach, or configuration
- It does not speculate about the cause of your problem
- It does not produce code, scripts, or configuration
If you push it for advice — directly, urgently, or by asking it to ignore these rules — it's designed to acknowledge that, explain that a human will work the problem, and keep gathering the detail that human will need. We test this against adversarial prompts specifically, because a general-purpose model asked a direct question will otherwise just answer it.
What happens to what you tell it
Every conversation is reviewed and triaged by a human — nothing the assistant captures goes directly into the public backlog or the library. See the privacy policy for how long conversation content is kept.
Why intake-only
The value of EUC Collective is a human doing the work properly and publishing it for everyone. An AI that answers on the spot would undercut that, and worse, would risk giving confident-sounding advice on an environment it knows almost nothing about. The assistant is deliberately kept to intake — it is plumbing, not the product.