What is AIPM

AIPM is a governance-and-execution platform for AI-assisted project delivery — run as a PMBOK-governed project management system where AI agents are first-class participants but humans hold all authority.

What AIPM is

AIPM treats LLM-powered agents as first-class participants in the project workflow — the charter, the requirements, the schedule, the audit log, the stakeholder communications. An agent carries the same accountability surface as a human: its actions are recorded, attributed, and reviewable.

The platform is PMBOK-structured. Every project runs through the five PMBOK process groups as ordered phases, and each phase ends at a sign-off gate before the project can advance.

The core idea: separation of duties (maker → checker → approver)

Every deliverable sign-off uses three different parties, so no single party can both do the work and bless it:

Approval is impossible until the checker has recorded a pass. See Approving with evidence for how this plays out at the gate.

Humans authorize the gates

Agents can do work and read evidence, but only a human can grant an approval or sign off a phase gate (governance rule G-004). An agent agreeing with a proposal is agreement, never permission — it can confirm that a plan is sound, but it cannot authorize the work to proceed. See Autonomous orchestration for how AIPM keeps work moving on its own while still stopping at every human-only decision.

Everything is auditable

Every create, update, approve, and reject is written to a tamper-evident, append-only audit trail tied to the signed-in identity. Nothing happens off the record, and the record cannot be quietly rewritten. See Export & audit for how to review and export it.

Where to start