elmo

What a pilot actually involves

Start from the report you need, not from the feature list

The useful first question is not which features you want. It is: at the end of this, what does somebody have to be able to prove, and to whom. A regulator asking who completed the annual training. A customer asking whether the engineers on their account are certified. A manager asking why half the team has not started.

Tell us which of those you are actually answering, and the rest of the scope follows from it — including whether elmo is the right thing at all, which we would rather establish in week one than in month six.

What we need from you

Less than you would expect, and none of it is a commitment:

What the two weeks look like

The platform installed on your cluster, your directory connected, one course deployed behind its own hostname and gated, a class opened with real dates and real people, and the reports produced from actual use rather than seeded data.

Then the part that matters: we sit down with the report and go through what each figure counts and what it does not, against the obligation you named in week one. If the answer is that it does not carry that obligation, that is a finding, not a failure — and you will have it before you have bought anything.

What you get at the end

A written account of what was built, what it measured, what it cost to run, and the current list of limits. If you stop there, you keep the document and the cluster stays yours. Nothing about a pilot is designed to leave you somewhere it is awkward to leave from.