Case studies

The work we can actually show you.

This list is short on purpose. A case study goes here only when the build is documented, the numbers trace to a dated run, and nothing on the page needs a reader to take our word for it. Everything else stays off.

The evidence standard

What is a SummitCited case study?

An engagement written to the same rules as our audits: every number traced to a dated run, every raw answer kept on file, no guaranteed rankings, and nothing claimed that we can’t show. Our own hand checks are labelled as such, never scored as runs. No client revenue or lead counts, because we don’t audit their books. No ranking promise, because AI answers vary run to run. No testimonial we haven’t been given, and no name we haven’t been cleared to use. What’s left is what we built, what the instrument read before, and what it reads at each dated run since, including the months it reads nothing new.

Documented engagements

One entry, held to the standard.

Irrigation & sprinkler service · Utah County

Smart Sprinkler Solutions

We fixed the site. ChatGPT went from naming the company in 22% of answers to 58%.

13 of 6022% · ChatGPT, 20 August · twelve buyer questions, five samples each, the day before the rebuilt site went live
35 of 6058% · ChatGPT, 17 September · the same twelve questions at the same depth, scored by the same code
Change detectedthe one like-for-like pair on file, 28 days apart; both engines' verdict lines, word for word, are on the case study

What moved ChatGPT: we fixed the site. A single-page app replaced by a nineteen-page site built from a verified data file, live 21 August, twenty pages now. The case study carries the verdict lines word for word and every other run on file at its depth, the errored ones included.

Read the full case study →

Engagement details are published with client agreement. Where consent to name outcomes isn't on file, the outcome isn't described — the build and the measurement are.

Claim my free audit