Unify. Understand. Act. Then again.

A single good decision is a moment. An organization needs the moment to repeat — every cycle, every team, without the discipline degrading each time it happens. Priorities.ai runs one loop, three stages, on a rhythm you set.

The loop, in one pass

Unify brings every candidate — new requests, the commitments already in flight, the people who have a stake — into one record, judged by the same criteria. Understand turns that record into a comparison anyone can follow: what outranks what, where two commitments compete for the same people or budget, what a change would actually cost. Act publishes the result, records why, and hands it to the teams doing the work — with a baseline the next signal gets checked against.

Then the loop runs again. A published decision does not close the record; it becomes the baseline the next Unify starts from.

     ┌─────────┐        ┌─────────────┐        ┌───────┐
     │  UNIFY  │ ─────▶ │ UNDERSTAND  │ ─────▶ │  ACT  │
     └─────────┘        └─────────────┘        └───────┘
          ▲                                         │
          └─────────────────────────────────────────┘
            the published decision is the baseline
                    for the next pass

01 — Unify

Before anything can be compared, it has to be in the same place, described the same way. Unify is where the Frame is set: who has a voice in this round, which items are eligible, and what criteria apply to all of them equally.

Unify collects:

Requests and commitments — new candidates and the work already promised, in one pool
Participants — who contributes judgment, who holds authority, who is only informed
Criteria — the dimensions of value that will apply to every item in the round, agreed before anyone scores anything

Without this step, “prioritization” is whichever list arrived most recently, compared against nothing in particular.

02 — Understand

With everything unified, Understand is where the organization sees how it actually compares — not by gut feel, but against the criteria set in the step before. The Cascade carries the comparison up and down the organization: leadership’s context flows down, teams’ ground truth flows up, and both inform the result.

Understand surfaces:

Where items rank, against the agreed criteria and weights
Where commitments conflict — the same people, budget or time claimed twice
What a dependency requires — a typed relationship graph, not a flat list, so a hard dependency or an all-or-nothing bloc is visible before it becomes a surprise mid-quarter

This is the step most tools skip past — they show a ranked list without showing the disagreement it papers over.

03 — Act

Act is publishing the decision — and keeping it accountable to what comes after. The result becomes the Plan of Record: what was decided, how, and why, with decision traceability attached so the rationale is queryable long after the meeting that produced it.

Act includes:

Publishing the ranked, agreed result to the teams executing it
Recording the decision — criteria, values, session, governance event — as a typed, permanent entry
Checking work in progress against what was decided, so drift is visible while it is still cheap to correct

Acting on a decision does not end the loop. It sets the baseline the next Unify measures against — so a competitive move, a market shift, a failed assumption becomes a question you can actually answer: is this a legitimate reason to change course, or a departure from a commitment that still holds?

Why it has to loop

A method run once is a workshop. A method that runs every cycle, on the same record, is infrastructure. Unify, Understand, Act is not three separate features — it is the shape every defensible priority takes, repeated on a rhythm the organization can sustain.