The quiet period is the cheap period
Between announcement and standards manual, preparation is at its cheapest: the evidence that takes longest is the evidence every framework already agrees on — and waiting means assembling it against a survey date instead of a calendar.
The Joint Commission announced its Responsible Use of AI in Healthcare certification on 1 June 2026. As of this issue, the standards manual is not public. Most organisations have read that gap as permission to wait.
That is the expensive reading.
Here is the mechanism. The evidence a certification asks for divides into two kinds: the kind you can produce in a week, and the kind that can only accumulate. A committee charter is a week. A year of quarterly performance attestations is a year — there is no purchase order that compresses it. When the manual publishes and a survey date exists, the week-kind evidence gets written in a hurry and the accumulating kind simply does not exist, because it could not be backdated. The organisations that look prepared at the first surveys will be the ones whose monitoring record started before anyone required it.
And the accumulating evidence is not a mystery, even without the manual. The Joint Commission’s September 2025 guidance with CHAI, and CHAI’s own governance playbooks from May 2026, agree on the shape: an inventory of what you run, a governance body that owns it, performance monitored against a baseline, bias examined, incidents reported, people trained. Whatever the final standards text says, it is not going to ask for less than the publicly stated consensus of its own named collaborators. That sentence is inference, and labelled as such — but it is the kind of inference a prudent operator plans on.
So the quiet period prices preparation at its floor. No survey date, no consultant surge pricing, no board deadline — just the option to start the clocks that cannot be started later. Register the inventory. Charter the committee. Lock a baseline for the highest-risk model you run and attest it once a quarter. Each of those is boring this month. Each becomes unbuyable at any price in the month a survey is scheduled.
The organisations that treat announcements as the deadline get eighteen months of accumulated evidence for the cost of a standing agenda item. The ones that treat the manual as the starting gun get a sprint, against a date, on the one kind of evidence that does not sprint.
The quiet period is not the pause before the work. It is the work, at the only discount it will ever be offered.
One operational argument a week
The Operations Edge lands each Monday: a hook, one thing to use before lunch, and the full argument here in the archive. No vendor sponsorship, ever.
The instruments behind the writing
Every framework in the series is published as a working file: registers, protocols, audit rubrics and unit-economics models, sized to be used rather than admired.
See the toolkits The library← Your downtime plan does not cover degradation Your AI arrived in a release note, not a purchase order →
Published under the Institute's editorial standard.
Author: Neel Chauhan, MD MBA, physician-executive and founder of the Healthcare AI Institute. Last reviewed against the standard on 2026-08-21.
Argued from the published record: the Joint Commission's announcement of 1 June 2026, its joint guidance with CHAI of September 2025, and the CHAI governance playbooks of 28 May 2026, all read in full. Where the Joint Commission has not published something, this issue says so and labels what follows as inference.
The Institute accepts no vendor sponsorship, holds no vendor equity and takes no referral fees.