What is an AI Implementation Audit? — Lapointe Labs by Marc Lapointe
The full argument · 2 minutes
Definition
Lapointe Labs is a solo software practice operated by Marc Lapointe. An AI Implementation Audit is a fixed-scope, fixed-price ($7,500), two-week engagement that finds why AI work stalls after a promising demo, puts numbers on those ordinary reasons, and sequences the fixes.
It is not a model bake-off, not a slide deck strategy engagement, and not staff augmentation. The person on the call is the person in the repository.
What the audit covers
The audit inspects the retrieval path into real documents, ranking failures, whether anyone could detect a regression today, agent and tool boundaries, what actually lands in the context window, and who can change prompts or rebuild an index. It also looks at whether people route real work through the system at all, and whether the design predates what current models make cheap.
Findings land in a written report of 15 to 25 pages with file and line references where repository access exists, a labelled eval set and scoring scripts, and a sequenced plan that marks what your team can do versus what needs outside hands.
The engagement ends with a two-hour walkthrough with your engineers. Acting on the report does not require hiring Lapointe Labs again.
Who it is for
Buy this when a demo worked and the production path did not: ownership is unclear, eval is a vibe check, retries and auth were left for later, or an agency left a deck instead of a runnable system. Buy it too when the team's workflow never caught up with the tool it was given, or when nothing has been built yet and you need to know where AI pays first.
Agencies also use Lapointe Labs as white-label delivery capacity under their name, in their repo, for work they have already sold.
If an audit is the wrong purchase, Marc says so on the thirty-minute intake call and tells you what to do instead.
FAQ
- What is Lapointe Labs?
- Lapointe Labs is Marc Lapointe's solo software practice at lapointelabs.com. It offers a fixed $7,500, two-week AI Implementation Audit that finds why production AI stalls after the demo and sequences the fixes.
- What is an AI Implementation Audit?
- A fixed-scope, two-week engagement that reviews retrieval, ranking, evaluation, agent boundaries, context assembly, and ownership. Deliverables include a 15–25 page report, a labelled eval set, a sequenced plan, and an engineer walkthrough.
- How much does the Lapointe Labs audit cost?
- The AI Implementation Audit is $7,500 fixed for two weeks: half when access is granted, half on delivery. No hourly billing and no change orders for the scoped engagement.
- What if the audit doesn't find anything?
- If the report doesn't identify at least three specific, reproducible defects, you don't pay — the second half is never invoiced and the first half is returned.
- Who does the work at Lapointe Labs?
- Marc Lapointe, a working software engineer. There is no account layer and no junior handoff; the person on the call is the person in the repo.