The Mycelium Group is an AI security and governance practice for Australian boards and regulated industries — built on three decades of diagnostic work on the network beneath high-stakes change. Three connected practices: Board AI Readiness (free 30-minute diagnostic), AI Security Health Check (4-week, board-ready scorecard), and Managed AI Agents (purpose-built departmental builds).
We help leaders determine whether further AI investment will convert into adoption — or whether the operating conditions beneath the program need to be fixed first.
Most organisations respond to stalled AI adoption with more: more tools, more training, more mandate. That response treats the symptom. If the operating conditions beneath the program are weak, additional investment does not solve the problem — it scales the waste. The network beneath the structure is the cause. In every sector. At every scale.
AI programs fail when five organisational conditions are absent. These are not cultural attitudes — they are structural operating conditions. The diagnostic scores your organisation across all five using executive interviews, frontline input, governance review, and artefact analysis — mapping precisely where adoption will fail and why, before further investment is committed.
People feel safe to experiment, fail, and report honestly before issues become political.
Determines whether delivery risk is visible early — or hidden until cost, delay, and adoption failure are already locked in.
What works in one team reaches other teams before adoption fragments.
Determines whether the organisation scales what works — or rebuilds the same solutions in isolation at ongoing cost.
Frontline teams can adapt AI tools to how they actually work, within clear governance boundaries.
Determines whether the program reaches operational reality — or stays a compliance exercise that bypasses real work.
People are rewarded for genuine use and outcomes, not reported activity.
Determines whether adoption data reflects reality — or reflects what people know leadership wants to see.
Leadership receives accurate data on what is and is not working, early enough to act.
Determines whether leadership can intervene before failure is locked in — or learns of adoption collapse after the spend is committed.
The gap between what leadership perceives and what the frontline experiences is not a people problem. It is a network health problem. It is precisely measurable — and it is the most reliable predictor of whether an AI program will hold.
gap between executive and frontline perception of AI readiness — in the same organisations
If adoption conditions are weak, additional AI spend does not solve the problem. It scales the waste.
The standard response to a stalled AI program is more: more tools, more licences, more training, more mandate. That response assumes the problem is insufficient input. It is usually insufficient operating conditions.
The diagnostic tests whether the next tranche of AI investment will convert into genuine adoption — or add to the sunk cost. It is designed as a decision gate, not a consulting engagement.
Every engagement begins with a diagnostic. No engagement proceeds without one. If the diagnostic does not support further work, we say so before a proposal is written.
The Organisational Ecology Model is the methodology under both diagnostics. The five conditions assessed are constant. The application is calibrated to each specialism.
Where every Mycelium engagement begins. A diagnostic, a workshop, an assessment — for boards now formally accountable for AI.
Where does your AI risk actually sit? A four-week board-ready diagnostic.
Mission alignment, workflow mapping, build feasibility. The front door to every Mycelium agent build.
Departmental agents, built against the mission, governed by the people who use them.
Every Mycelium engagement begins with a diagnostic. Where the findings support further work, governance, enablement, and operation follow — in that order.
Australian boards are now formally accountable for AI in a way they were not eighteen months ago. AICD and HTI have published five major pieces of director-duty guidance. APRA CPS 230 is in force. AS ISO/IEC 42001 has been adopted as an Australian standard. The board's exposure has changed.
Board AI Readiness is the front door to every Mycelium engagement. A free 10-question diagnostic, a 3-hour facilitated workshop, and a 3-week readiness assessment — with each rung funding the discovery for the next. Most boards never need to go past rung two.
A four-week AI Security Health Check across 14 structured dimensions — Govern, Protect, Operate, Enable. Aligned to the Australian Government's six Essential AI Practices, plus eight additional control areas regulators are likely to formalise next.
Two deliverables. One clear picture. An Executive Scorecard ready to table at your next board meeting, and a Detailed Findings Report with a prioritised remediation roadmap. Board-ready before 10 December 2026.
Three purpose-built AI agents are available to commission now. Each is built against a department's specific mission, governed by the people who use it, and released only after executive sign-off. No off-the-shelf compromise.
Phase 1 demos are available for all three — a working prototype on synthetic data, no production access required, in 4 weeks. The right way to show leadership what's possible before a full build is commissioned.
Organisations engage when major initiatives are failing at implementation — where strategy is clear at the top and dissolves before it reaches the front line. The operating conditions beneath the structure are the cause. In every sector. At every scale.
Coherence is fracturing. The operating conditions that worked at 80 people no longer function at 300. The diagnostic shows where they need to be deliberately rebuilt — before the next growth phase, acquisition, or strategic shift.
Strategy is sound. Investment is committed. The program is not landing. Stage 1 identifies precisely where the operating conditions are failing to carry the change — and what it will take to restore them before further spend is made.
Organisations engage at different stages of the same problem. The diagnostic conversation is the same regardless of where you are — it determines what is causing the issue, which stage applies, and whether we are the right firm to address it.
Stage 1 confirms readiness before capital is allocated. If conditions are not present, you will know precisely what it will take to create them — and what it will cost to proceed without doing so.
Assess whether the organisation can carry the investment before it is committed. Request a Diagnostic Conversation →
The program is live, the technology is deployed, adoption is not following. Stage 1 identifies the specific operating conditions responsible. Most stalls are diagnosable within weeks.
Diagnose why adoption is not holding before investing in more tools or mandate. Request a Diagnostic Conversation →
Repeating the program without changing the conditions produces the same result. Stage 1 determines what failed and why. Stages 2 and 3 are offered only when findings support a viable path forward.
Determine what broke and whether the conditions exist to restart. Request a Diagnostic Conversation →
Additional investment in an unready network extends the loss. Stage 1 confirms whether the next tranche of spend will produce genuine use — or add to the sunk cost.
Test whether additional spend will convert to adoption or compound the problem. Request a Diagnostic Conversation →
Sixty minutes. We assess the presenting problem, determine the right entry point, and establish whether our approach is the right one for your situation.
No pitch. No proposal until it makes sense.
Funding AI programs the organisation cannot absorb
Scaling low adoption with further spend
Receiving false-positive program status from filtered reporting
Discovering adoption failure after the investment is committed