Domain — Identity Governance & Security

Your identity programme governs your employees. It does not govern everything else with a credential.

The lifecycle a service account never had
  1. 01

    Created

    Usually by whoever needed it, with no record of why.

  2. 02

    Owned

    Assigned to a named person who answers for it.

  3. 03

    Recertified

    Its access reviewed on a cycle, like an employee's.

  4. 04

    Retired

    Revoked when its purpose ends. The step that is usually missing.

01

Know every non-human identity in your estate, and who owns it.

Every human identity has a lifecycle. Service accounts, API keys and integrations are created by whoever needed them, and there is no equivalent of leaving. When a person exits, the credentials they stood up remain — with their authority and without their accountability.

We enumerate the population across directory, cloud, database and integration layers, and record a named accountable owner for each. Where no owner can be identified, that finding is itself the result.

02

Bring service accounts inside joiner-mover-leaver.

Joiner-mover-leaver is the control your institution already trusts. It was written for employment events, so a credential created for an integration never joins, never moves and never leaves.

We define the equivalent events for non-human identities — creation with an owner and a purpose, change on system or ownership transfer, retirement on decommission — and set out how existing processes absorb them without new tooling.

03

Close the recertification blind spot.

Access recertification is the primary compensating control in every regulated institution. Non-human identities have no manager to attest and no review cadence. The control auditors rely on does not cover the fastest-growing population.

We establish what would have to be true for these identities to be attestable — owner, purpose, expected behaviour, evidence of use — and where current records fall short of that.

04

Keep segregation of duties intact when the actor is software.

Segregation of duties assumes two actors. A single service or agent that can both initiate and approve collapses that into one identity, without triggering any control designed to detect it.

Software also holds the union of every privilege it might need and, unlike a person, exercises all of it continuously. Excessive privilege stops being latent risk and becomes active behaviour. We identify where that has happened in the processes your regulator examines.

The offer

The Machine and Service Account Exposure Baseline

A fixed-scope, fixed-fee assessment. Three weeks. Read-only, no agents, no production change. Ends in a document your risk committee can act on.

Take the AI Readiness Assessment

Scope and fee are agreed before work begins.

Two halves of one argument

Non-human identity has been a quiet problem for a decade. AI is what makes it loud. Every agent an institution deploys is another identity with standing authority, no owner, and no review cycle — built on the same foundations that already govern your service accounts poorly.

AI Governance & Security