The Method
A four-gate methodology that produces its own audit trail.
Project Management Artificial Intelligence System (PMAIS) applies the discipline of safety-critical engineering to project delivery: staged gates, defined entry criteria and written evidence at every transition. Nothing advances without it.
PMAIS has two layers. The Core is the base methodology: four phases, four governance gates, applicable to any project in any sector — this is what LeoxPM runs directly. AIS-IMG is the Core's specialization for deploying AI systems into regulated or high-risk environments: it adds a fifth gate and nine implementation stages for continuous, evidence-backed governance — this is what RoveonPM runs. What follows on this page is the Core's four-gate model; AIS-IMG's five-gate model, the fullest expression of PMAIS's governance approach, has its own page.
The four gates
Initiation, Planning, Execution, Closure.
A project that cannot clear a gate does not proceed — and that decision is documented too.
PMAIS applies to any project, in any industry — regulated or not. Gate enforcement scales with risk, not with sector: a general project run through LeoxPM uses lighter, mostly-advisory checks at each gate, since a human is directing the work throughout. A project that deploys an AI system into production — run through RoveonPM, PMAIS's AIS-IMG specialization — enforces gates strictly, with a named human reviewer required before any AI-drafted output is released, because the AI itself is doing more of the work unattended.
- 1
Initiation & Business Feasibility
Gate 1 exit criteria: approved Project Charter, ROI/value case above threshold, Automation & Oversight Plan drafted, high-level data-privacy assessment completed.
- 2
Planning & AI Pipeline Configuration
Gate 2 exit criteria: signed Project Management Plan, AI Agent Configuration Record, baseline schedule and RAID log established.
- 3
Execution & AI-Assisted Delivery
Gate 3 exit criteria: delivery milestones met, AI-draft acceptance/correction rate within tolerance, no unresolved data-governance or compliance flags.
- 4
Closure, Handover & Capacity Realization
Gate 4 exit criteria: final acceptance/sign-off recorded, lessons-learned and Capacity Realization Report completed; where the project produced or touches an AI system that continues into production, a handover to RoveonPM/AIS-IMG governance is recorded.
Inside each gate
What actually happens.
Initiation & Business Feasibility
Validate business justification, technical/data feasibility, regulatory exposure and risk profile before resourcing the project, and set the Automation & Oversight Plan. Gate 1 exit criteria: approved Project Charter, ROI/value case above threshold, Automation & Oversight Plan drafted, high-level data-privacy assessment completed.
Planning & AI Pipeline Configuration
Finalize the schedule baseline, RAID structure and stakeholder plan, and configure which AI agent roles are active. Gate 2 exit criteria: signed Project Management Plan, AI Agent Configuration Record, baseline schedule and RAID log established.
Execution & AI-Assisted Delivery
Run the project, with the AI Execution Layer producing periodic drafts and the Human Governance Layer reviewing, correcting and approving them. Gate 3 exit criteria: delivery milestones met, AI-draft acceptance/correction rate within tolerance, no unresolved data-governance or compliance flags.
Closure, Handover & Capacity Realization
Close the project, hand over any operating artifacts, and formally measure the capacity freed. Gate 4 exit criteria: final acceptance/sign-off recorded, lessons-learned and Capacity Realization Report completed; where the project produced or touches an AI system that continues into production, a handover to RoveonPM/AIS-IMG governance is recorded.
Principles
Three commitments that shape every engagement.
Evidence before enthusiasm
Every claim about performance, fairness or safety must be reproducible from an artifact, not from a vendor deck.
Reversibility by design
Any system we put into production can be withdrawn without halting the business process it supports.
Accountability is named
Each control, decision and exception has a person attached to it, recorded at the time the decision was made.
Applied
PMAIS is the spine of two products.
LeoxPM runs the 4-gate core directly — general project management for any team, any sector. RoveonPM adds a 5th gate and nine delivery stages (AIS-IMG) purpose-built for governing AI systems from pilot through production.