Receipts
What moved, with the caveats attached.
We publish client outcomes the way we'd want to read them: real numbers, dated, with the confounders stated. Clients stay anonymous. When something didn't move, we say so. This page grows one receipt at a time.
Receipt 001 · Developer-tools company, open-source core · 2026
Cited four times as often, two weeks after the work merged.
The situation
A developer-tools company with an open-source core. Strong product, real adoption, and a heavy public GitHub footprint that nobody had touched in years. When a buyer asked an AI assistant who does what they do, they barely showed up: 1.25% of category answers cited their own domain. The baseline scan said the cheapest fixes weren't on their website at all.
What we actually did
AI assistants read READMEs as product documentation, not as onboarding pages. Theirs described how to install the code and said almost nothing about what the product is or who it's for. So the work was surgical:
- Wrote one canonical sentence saying what the product is and who it's for
- Added a real comparison table and an FAQ to the main product pages
- Put that same canonical sentence at the top of the README in 16 public repositories
- Shipped it as 14 pull requests their team reviewed and merged, in roughly two weeks
No site rebuild. No paid traffic. No new content program. The measurement said where the leverage was, and the work went there and nowhere else.
What moved, and how fast
1.25% → 5.8%
share of category answers citing their own domain
It stepped the day the pull requests merged and held there for two straight weeks. Not a slow climb: a step, the week the work landed.
Their own tracker agreed, independently
They run a third-party AI-visibility tool alongside ours. Neither was tuned to the other, and both moved in the same week:
- Visibility score rank up 15 places → #10
- Share-of-voice rank up 15 places → #10
- Average position 8.5 → 4.1
The point is the speed, not the size of the number. AI visibility is not SEO's two-to-three-month flywheel: when the fix is aimed at what assistants actually read, movement shows up in weeks.
The caveat
Their site team shipped linking work in the same window, so we call this strong evidence rather than proof. A receipt without its confounders is marketing; this page doesn't do that.
How we count
Baseline scan before the work, the same question set after, on Collimer, our AI-visibility agent. Where the client runs their own tracker, we report it alongside ours. Outcomes are reported in one of five states, from "verified moved" to "moved against"; nothing is collapsed into a thumbs-up.
How this happens for you
This receipt is the first two weeks of the engagement we sell as the AI GTM Sprint: 90 days, fixed price. We baseline what AI says about you on Collimer, fix what the measurement says is cheapest to move, and verify it landed. Start with the free scan if you want to see your own baseline first.