An engagement is a second read: an independent re-review of one number, one list or one standard, by someone qualified to disagree with it, before you spend against it. It is a fixed-fee assessment with a defined end, not a retainer.
How one starts. Write with the figure or the list you are being asked to act on. A redacted list is enough; the first look needs no member data and no claims, and I will tell you whether there is anything worth doing before either of us commits to a scope. That first look costs nothing. If there is something worth doing, we agree what is in scope, what data it needs, and what it will cost, in writing, before any work begins.
What it produces. Assessments run four to six weeks and end in a findings report and a sequenced plan you can take to a governance forum. Every number in the report carries its provenance: measured from your data, fitted, taken from a published source, set by statute, or assumed, and it says which. Every threshold resolves to a declared source. The limitations section is generated from the same flags the instruments raise on screen, not written as boilerplate, so the report states its own limits, which is what lets it survive being checked by the next person who reads it.
You also keep the engagement itself. The work lives in a file you hold, and the instruments that produced it are the same ones published on this site, free and open, with nothing held back. Six months after the engagement ends you can reopen the file and reproduce every figure. That is the point of building a practice as a set of instruments rather than a set of decks.
What it costs. A fixed fee, set at scoping and stated before the work starts. The fee depends on the scope, not on the size of the finding, so there is no incentive on this side to inflate a number. Implementation support follows if it is useful, and is scoped and priced separately.
What it does not do. It does not make clinical calls; the arithmetic never decides an authorization, an hours reduction or a denial, and the report prints what the numbers cannot carry. It does not price clinical harm, anywhere. It does not count deterrence as savings without a way to evidence it. And it does not blend the instruments into one score, because they are built to disagree with each other in useful ways.
What happens to your data. Nothing leaves your browser. The instruments run entirely on the page, with no server, no database and no outbound network, enforced by policy rather than promised in a footer. Client data stays in your custody throughout; the practice works from what you choose to share, and the engagement file is yours to keep or destroy.