Home/Professional Workflows/For Project Managers

Pilot Human-facilitated

AI-Augmented Project Manager

You have used AI on project work. It drafted the status report, summarised the risk log, tidied the charter. Then you went back to a project system that had not changed at all. This pilot is about changing the system.

Apply to join the pilot Register for the Benchmark Compare delivery routes

The problem, precisely

AI arrived in project management as a set of task accelerators. It writes faster than you do, summarises reliably, and produces documentation that looks right. So it got adopted quickly: at the task layer, by individuals, without governance.

What did not happen is integration. There is usually no defined point at which AI enters the project workflow, no allocation of what a project manager must own versus what a model may draft, no verification step that catches a plausible-but-wrong risk assessment before it reaches a steering group, and no record of any of it. When someone asks how AI is being used on the programme, the honest answer is often "variously, by different people, in ways we have not documented."

That is not just a productivity gap. It is inconsistent quality, a thin audit trail, and no defensible answer when a steering group asks how AI was used on the programme. A policy document telling people to be careful does not fix it.

What you build in the pilot

You will leave with one documented project workflow, built on your live project, with defined AI entry points, review steps, decision ownership and stop conditions — specific to your organisation's standards, not a generic template.

The point is not to use AI on more tasks. It is to build an AI-native project workflow — you and AI operating as one governed unit — so that you become an AI-augmented project manager able to govern more of the delivery yourself. What "AI-native" and "AI-augmented" mean → This is a first paid pilot, not a fully developed product family.

  • A defined AI entry point in the project lifecycle, rather than ad hoc use wherever someone thinks of it.
  • An explicit allocation of what a project manager owns, what AI may draft, and what still requires a specialist or a stakeholder decision.
  • Verification steps designed for the specific failure modes of AI in project work — confident wrong estimates, invented dependencies, risk registers that read well and miss the actual risk.
  • Escalation triggers that stop the workflow when inputs are insufficient, when accountable judgment is required, or when a claim would outrun its evidence.
  • A governance position you can defend to a PMO, an auditor or a steering group, because the workflow is documented and its controls are explicit.
  • An evidence record of workflow operation: judgment applied, AI contribution, validation decisions, output quality against standard.

Diagnostic overlay

The project management overlay adds these constructs to the common core:

  • Integration into the project lifecycle
  • Governance position and defensibility
  • Decision traceability
  • Stakeholder-facing quality control
  • Documentation and reporting integrity
Register for the profile

Who this is for

Project, programme and delivery managers with a live project to work on during the cohort. It is not suitable if you have no current project: the method requires real work to test against at Stage 7, Test.

Honesty about the pilot

What "Pilot" means, and what we are not claiming.

  • The design method exists, is documented, and we use it ourselves.
  • Places are limited and allocated by application, so the work can be supported properly. A live project is required: the method is tested against real work at Stage 7.
  • The workflow you produce belongs to you.
  • We do not yet have client outcome evidence — the cohorts are running now — so we make no claims about delivery improvement, time saved or audits passed.

How we classify evidence, and what stands behind everything above, is on the evidence policy →

Boundaries

Decisions the workflow will route to the accountable specialist.

Project work carries contractual, financial and safety consequences. Designing where the workflow stops is part of the governance value, not a disclaimer.

Contractual and commercial decisions

Contract interpretation, claims, variations and commercial positions escalate to qualified legal and commercial specialists. The workflow drafts nothing that binds anyone.

Safety, regulatory and compliance matters

Safety cases, regulatory submissions and compliance determinations go to the accountable qualified professional. Always.

Confidential and client-owned project data

Project data boundaries, client confidentiality and information-security constraints are designed into the workflow as constraints, not handled by reminder.

Estimates and forecasts presented as fact

Any figure that will inform a decision requires a stated basis and human validation. A confident number from a model is not an estimate.

Applications

Apply to join the pilot cohort.

Applications are read by a person. We are looking for project managers with a live project, some existing AI use, and an actual governance problem to solve, not for the most impressive CV.

Apply now →

Your project information

What you tell us in this form and during the cohort stays yours. Your project data, your organisation's information and the workflow you build remain under your control.

If we ever want to use anything from your work as a product example or case study, we ask you first, in writing, and we anonymise unless you tell us otherwise. Nothing is published without consent. Privacy notice →

Not ready to apply

Register for the benchmark with the project management overlay. When open, it will give you a baseline and tell you honestly if the pilot is the wrong next step. Register for the Benchmark →