Turn transformation evidence into the next move.

The Transformation Agent is an AI reasoning layer over organizational transformation evidence and history. It identifies attention areas, investigates gaps and prepares governed recommendations for human approval. It works inside Transformics, on top of your Transformation Performance Index and your organization's transformation record.

Three things. The Transformation Chief of Staff gives a concise readout of up to five items that need attention. The Gap Investigator looks into one gap: what is known, competing explanations with the evidence for and against each, a recommended next step and what is still unknown. Governed Agent Actions prepare a concrete proposal, such as a suggested initiative or a decision brief, for a person to review.

No. The Agent reasons on top of scores that are already computed. It never changes a score, a weight, the methodology or an evidence item, so your TPI is produced the same way whether the Agent is on or off.

It prepares them; it does not carry them out. A prepared action waits in a review queue until a person with approval rights reviews it, edits it if needed and approves or rejects it. An approved initiative is added to the Steering Board as a suggested initiative, and accepting it is a separate step. A decision brief is a draft for the people who decide; the Agent never records a decision itself. It never writes to external tools such as Jira, Slack or email.

It starts from the record your organization has already built in Transformics: current state, previous changes, decisions, initiatives, evidence and the practices your organization has validated. Each item carries its source and how current it is. The Agent's own working notes stay provisional and only become part of the organizational record if a person confirms them.

Only what your organization already has in Transformics: scores and pillars, quarter-over-quarter movement, cross-module findings, evidence, initiatives, governance signals and transformation memory. It receives an allowlisted, aggregated slice of that data, without owners, emails or free text that could identify a person. It does not browse the web or connect to other systems.

No. The Agent works on aggregated, team-level data and never below a minimum group size. It is not intended for individual performance evaluation.

Its reasoning is read-only, and every recommendation is checked against your data before you see it: references that do not exist and unsupported claims are removed. Every run is recorded in the platform's audit trail, with only a hash of the prompt stored. Proposed actions need a person's approval, and the Agent is switched on per workspace.

Reasoning (the Chief of Staff and the Gap Investigator) is analysis you read: it changes nothing. Agent Actions are the next step: the Agent turns a recommendation into a concrete proposal that waits for a person to approve or reject it, and only an approved proposal enters your existing steering workflow.