Mycelium AI Agents — Agent 02 of 05

The PM Agent.

Built against the mission. Governed by the people who use it.

Status
Available now
Now accepting commissions
Investment
$35K–$180K
Fixed fee, three tiers, ex GST
Timeline
8–20 weeks
Five-phase managed build
The Proposition

Off-the-shelf AI doesn't know how your PMO works.

Generic copilots don't know your delivery cadence, your risk language, or what "done" means in your portfolio. They produce generic outputs — and the PMO ends up curating the output instead of running the work.

The PM Agent is built for one PMO function, against that function's mission, by a firm whose primary practice is diagnosing why AI fails. It synthesises status across projects, surfaces risk before it reaches the steering committee, drafts board-grade reporting, and hands every decision back to the humans who own it — with reasoning attached.

"Status compilation, board-pack drafting, and steering-committee prep are absorbing 30–40% of senior PM time. The PM Agent moves that work to a governed system, freeing senior PMs for delivery." The pattern across PMOs at scale

What it does

Three categories of work. One agent, governed.

— It Synthesises

The work that consumes senior PM time.

  • Cross-project status across the portfolio
  • Stand-up summaries and decision logs
  • Steering committee briefing packs
  • Board-grade reporting drafts
  • Stakeholder-specific narrative summaries
— It Surfaces

The risks that arrive too late to act on.

  • Risk signals from delivery data
  • Resource conflicts before they cascade
  • Slipping milestones at the earliest indicator
  • Stakeholder escalation patterns
  • Cross-project dependency breaks
— It Hands Off

Decisions to the humans who own them.

  • Decisions to the relevant decision-maker
  • Escalations to the PMO Director
  • Edge cases to the working group
  • Always with reasoning attached
  • Never makes the call itself
When it earns its keep

Three scenarios where the PM Agent changes the calculus.

01
Monday morning status synthesis
The recurring weekly burn

Senior PMs spend Monday morning collating updates from 8–15 project leads, normalising the data, and producing a portfolio status that the PMO Director can present at the 10am leadership stand-up.

The PM Agent ingests project data overnight, drafts the portfolio status, flags the three things the PMO Director needs to know first, and lands the draft in the PMO Director's inbox by 7am — with sources cited and confidence levels named.

Outcome: Senior PMs get Monday morning back. The Director walks into the stand-up with sharper context, not just a summary.

02
Steering committee preparation
The recurring cycle drain

Quarterly steering committee prep absorbs 40–60 hours of senior PMO time per cycle — pulling together project narratives, risk registers, financial summaries, and decision-required items into a coherent pack.

The PM Agent drafts the entire pack against the steering committee's standing agenda, with each section pre-populated from the underlying delivery data. The senior PMO team reviews, edits, and adds judgement — without writing the first draft.

Outcome: Steering committee prep moves from 40–60 hours to 8–12 hours. Time recovered goes back to delivery.

03
Risk surfacing before the steering committee
The escalation that arrives late

The most expensive PMO failure mode is risks that reach the steering committee already in red — when the time to act on them was three weeks earlier, when they were still amber.

The PM Agent monitors leading indicators across the portfolio — slipping commitments, resource gaps, dependency breaks, narrative drift in status updates — and surfaces the amber-to-red transition trajectory to the PMO Director the moment it begins, with the underlying signal visible.

Outcome: The PMO Director gets warning weeks earlier. Steering committee escalations become rare, not recurring.

How we build it

Every PM Agent build follows the same five-phase loop.

Most agent builds skip phase one (no mission alignment), compress phase two (assumed workflows, not observed ones), and have no equivalent of phases four and five (built and released without the people who will use it ever signing off).

The Mycelium build loop is slower than the market default. The slowness is the point.

01
2 weeks
Mission Alignment
Workshop with the PMO Director and 3–5 of the team. Output: Mission Alignment Document — what the function exists to do, what success looks like, what the agent must never do.
02
2 weeks
Workflow Mapping
Observation and process tracing of how the work actually happens — not how it is documented. Output: workflow map with the tasks, data, and human handoff points.
03
3–6 weeks
Build
Agent built against the workflow map and the Mission Alignment Document. Iterative, with the PMO working group reviewing each iteration. No dark builds.
04
2 weeks
Departmental Review
Parallel run alongside existing processes. The team uses the agent in production conditions, reviews accuracy and edge cases, and signs off before release.
05
1 week
Executive Release
Executive briefing covering scope, governance, review cadence, and performance measures. Released to broader use only after executive sign-off.
Investment

Three tiers, scoped to complexity.

Standard
$35K–$55K
8–10 weeks
Single workflow, single PMO, off-the-shelf integrations.
  • Status synthesis or steering prep — one workflow
  • Standard integrations (Microsoft, Atlassian, Asana)
  • Single-PMO scope
  • Five-phase build with executive sign-off
  • Handover documentation
Enterprise
$95K–$180K
14–20 weeks
Multi-workflow, multi-PMO or regulated environment.
  • Full PMO function coverage
  • Multi-PMO or regulated industry context
  • Compliance and audit-trail requirements
  • Extended departmental review (4 weeks)
  • Handover plus 90-day stabilisation

Fixed fee, ex GST. Optional Sustained Alignment Retainer ($3,500–$8,500/month) available post-release for quarterly mission re-alignment, drift monitoring, and governance review.

The governance promise

Three sign-offs. No exceptions.

Every PM Agent is released to broader use only after three sign-offs.

This is what separates a managed build from a deployed tool. The agent is not finished when the code is written. It is finished when the people who will use it have signed off on what it does, what it doesn't do, and how its performance will be reviewed.

01

The working group

The PMs and analysts who will use the agent daily sign off that it works as expected, handles edge cases responsibly, and aligns with the mission as scoped.

02

The PMO Director

The function owner signs off on scope, governance cadence, performance measures, and the explicit list of what the agent will and will not do.

03

The executive sponsor

The budget holder receives the executive briefing, confirms the scope and risk envelope, and authorises broader release.

The first conversation is diagnostic

No pitch. No proposal until it makes sense.

60 minutes with you and your executive sponsor. We assess whether a PM Agent build is the right move for your situation — and tell you if it is not.

Request a Diagnostic Conversation Download the PM Agent Brief
admin@themyceliumgroup.com.au   ·   +61 401 844 836