Our methodology
How the research is actually done.
Fast research earns its trust by showing its work. The method behind the deliverables: how we pick the peers we read, how we source every claim, how we tell you what we are sure of, and what we verify before it reaches you.
The comparator scan
We start by choosing the right peers. For a city, that means communities matched on the dimensions your question turns on: size, region, mandate, the specific service in play. We then read each one against the same set of questions, so the comparison is like-for-like and the gaps between you and the field are real.
Citation discipline
Every claim carries its source and the date we retrieved it. Public records change: a bylaw gets amended, a budget line moves, a program closes. A retrieval date tells you exactly how current a fact was when we pulled it, so you can rely on it or refresh it. A finding without a source does not make it into a deliverable.
Confidence tagging
Not every fact is equally solid, and we say which is which. We tag each claim as verified, reported, or estimated. Verified means we confirmed it against a primary source. Reported means a credible source states it and we have not independently confirmed it. Estimated means we derived it, and the basis is shown. A figure pulled straight from an audited financial statement is verified; the same figure quoted in a news article is reported; a per-capita rate we calculated from two public numbers is estimated. You always know how much weight a number can bear.
- Verified
- Confirmed against a primary source: an audited statement, a published bylaw, a dataset we pulled ourselves.
- Reported
- Stated by a credible source we did not independently confirm: a news report, an interview, a published vendor figure.
- Estimated
- Our own calculation or inference, with the basis shown: a per-capita rate we derived, a range we modelled.
Everything client-facing carries our sign-off
We read everything before you do: what matters, what to recommend, what is too thin to claim. Nothing carrying your name goes out without that judgment.
What AI-native means in practice
It means we use AI to extend what a small team can analyze, never to replace the analysis. The research engine reads more sources, scans more peers, and drafts faster than any team could on the same budget. Deciding what the evidence means, and what you should do about it, stays with us, every time. The speed is real; so is the line between the tool and the judgment.
Where our numbers come from
When we cite volume figures about our own delivery, the pages we assess, the peers we benchmark, the speed we work at, those describe our own pipeline. They are our internal measurements of our own work, not third-party-audited statistics. The deliverables themselves are the externally checkable proof.
Want to see the method on your question?
Tell us the decision the research has to support. We talk through how we would approach it and show you the method before you commit.
Start a conversation