Executive Team AI Exposure Dashboard — 11 C-Suite & Senior Leadership Roles × 175 Tasks
60% of all executive tasks sit in the Augment quadrant — high business case for AI, but current generative AI cannot reliably execute them without human oversight. This is the defining characteristic of a physical-infrastructure company: the value is created in Atoms, not Bits. AI can prepare, model, and draft — but cannot pour concrete, pull cable, or negotiate face-to-face with a Western Power program manager. The strategic question for GenusPlus is not "will AI replace our executives?" but "how fast can we shift augmentation-zone tasks into reliable AI-assisted workflows?"
GenusPlus's value chain is anchored in physical delivery — transmission line construction, substation builds, fibre rollouts, vegetation management. These are irreducibly physical activities. The C-suite's low Automate Now score (4.0%) reflects this: even where AI could theoretically help (document drafting, financial analysis), the consequence asymmetry is extreme — a wrong AI-generated variation claim on a $180M 330kV line project is not an "oops, redo." Secondary constraint is Institutional: ASX continuous disclosure, WHS regulations, and Australian award wage complexity create verification barriers (low V-scores across the board).
Click any card to view detailed task analysis in the Task Map tab.
The CFO (Damian Wright) and EGM Corporate Services (Mike Green) sit closest to the automation frontier — their tasks are more Bits-oriented (financial reporting, compliance, IT systems). The operational roles (George Lloyd, Kevin Arnold, Stewart Furness, David Fyfe) cluster in the high-IMPACT/low-FAVES zone — high business value locked behind physical delivery barriers. The CEO sits in the middle: strategic enough to be augmentable, but relationship-dependent enough to resist automation. Hasan Murad's Commercial role is the sleeper — high IMPACT but claims/dispute work requires deep domain expertise that AI cannot yet verify.
| Task | Category | IMPACT | FAVES | Quadrant | Freq% |
|---|
Tasks in the top-right (green zone) are ready to automate. Tasks in the top-left (orange zone) have strong business cases but AI cannot yet do them reliably — augment with human oversight. Bottom-right tasks are easy for AI but low priority. Bottom-left is the human fortress: leave alone. The overwhelming majority of GenusPlus executive tasks land in the top-left augment zone — a signature pattern for infrastructure companies where the value is in physical delivery and relationship judgment.
Across all roles, the V dimension (inverse consequence of error) scores lowest. This is characteristic of ASX-listed infrastructure: wrong outputs in regulatory filings, safety reports, contract claims, and financial statements carry severe penalties. V is the structural limiter — even where AI capability exists (high F, A, S), the consequence asymmetry prevents full automation. This is an Institutional constraint layered on top of the Atoms constraint.
All executives show headcount multipliers between 0.87x and 0.93x — meaning AI could theoretically compress each role's workload by only 7-13%. This is exceptionally low compared to knowledge-work benchmarks (where multipliers of 0.5-0.7x are common). The reason: physical delivery, relationship capital, and regulatory consequence create structural floors on human involvement. GenusPlus's executive team is well-insulated at the task level — the risk is not in C-suite displacement but in mid-tier operational and administrative roles below this analysis scope.
The deskilling index measures what proportion of each role's highest-value expert tasks fall in the Automate Now quadrant — indicating whether AI is consuming the complex work and leaving routine residue.
| Role | Deskilling Index | Interpretation |
|---|---|---|
| David Riches | 0.0 | Complex work remains with human — low deskilling risk |
| Damian Wright | 0.0 | Complex work remains with human — low deskilling risk |
| George Lloyd | 0.0 | Complex work remains with human — low deskilling risk |
| Strati Gregoriadis | 0.0 | Complex work remains with human — low deskilling risk |
| David Fyfe | 0.0 | Complex work remains with human — low deskilling risk |
| Hasan Murad | 0.0 | Complex work remains with human — low deskilling risk |
| Mike Green | 0.3 | AI absorbing some complex work |
| Kira McNeill | 0.0 | Complex work remains with human — low deskilling risk |
| Jane Carr | 0.0 | Complex work remains with human — low deskilling risk |
| Kevin Arnold | 0.0 | Complex work remains with human — low deskilling risk |
| Stewart Furness | 0.0 | Complex work remains with human — low deskilling risk |
Only Mike Green (EGM Corporate Services) shows any deskilling signal (0.3) — reflecting that finance/IT reporting tasks are being absorbed by AI tools. All operational and commercial leaders show zero deskilling — their expert judgment tasks (claims management, project negotiation, safety governance, industrial relations) remain firmly in the human fortress or augment zones. This is structurally protective for GenusPlus at the executive level.
With <30% Automate Now across all roles, the recommended portfolio allocation is: 10% Operator / 50% Fortress / 40% Escalate. GenusPlus should double down on its structural advantages (field delivery, safety governance, relationship capital) while strategically escalating commodity outputs up the value chain.
Limited application at C-suite level. Where it applies: CFO team can operationalise AI for financial reporting automation, variance analysis, and ASX announcement drafting. SHEQ team can deploy AI for safety analytics dashboards and regulatory change monitoring. Corporate Services can automate month-end close, AP/AR, and payroll compliance. These are the only roles with material Automate Now task clusters.
This is GenusPlus's primary defensive strategy. The moat is in: (1) David Riches' founder-operator network and client relationships that drive repeat business from BHP, Western Power, Ausgrid, FMG; (2) Hasan Murad's construction claims expertise — AI cannot prosecute or defend a $180M variation claim; (3) Jane Carr's enterprise bargaining and union negotiation capability; (4) George Lloyd and Kevin Arnold's ability to mobilise and manage field crews on remote transmission builds; (5) Kira McNeill's safety governance in high-risk environments. Invest in deepening these fortress capabilities — they compound with experience and are structurally un-automatable.
Where AI commodity outputs can be wrapped in higher-value services: (1) Use AI-generated financial models and tender pricing as inputs to executive judgment — the model output is commodity, the "David Riches says bid this at X margin" is transformation; (2) Use AI-drafted contract templates as starting points that Strati Gregoriadis's team customises with institutional knowledge; (3) Deploy AI safety analytics as inputs to Kira McNeill's risk governance — the dashboard is commodity, the safety culture is transformation; (4) Let AI prepare Board packs and investor materials that David Fyfe and Damian Wright then annotate with operational context that no AI has access to. The pattern: AI produces the draft, humans add the judgment, clients pay for the judgment.
| Role | AI Commodity Layer | Human Escalation Layer | Value Chain Position |
|---|---|---|---|
| David Riches — CEO | Market scan briefs, competitor analysis, draft strategy papers | Founder network, client trust, capital allocation judgment | Transformation |
| Damian Wright — CFO | Automated reporting, variance analysis, compliance checklists | Capital structure decisions, banking relationships, M&A valuation judgment | Service → Experience |
| George Lloyd — EGM | Project status dashboards, scheduling optimisation | Field delivery leadership, client relationships in mining/energy | Experience |
| Strati Gregoriadis — GC | Contract template drafting, regulatory monitoring, due diligence checklists | Dispute strategy, Board governance judgment, ethical culture | Experience → Transformation |
| David Fyfe — COO | Operational dashboards, KPI tracking, resource scheduling tools | Utility client relationships, workforce mobilisation, escalation management | Experience |
| Hasan Murad — Commercial | Standard form contracts, claim quantum calculations, schedule analysis | Dispute prosecution/defence, regulator negotiation, expert witness credibility | Transformation |
| Mike Green — Corp Services | AP/AR automation, payroll processing, IT ticket resolution | Corporate policy design, vendor strategy, digital transformation leadership | Branded Product → Service |
| Kira McNeill — SHEQ | Safety analytics, audit checklists, regulatory change alerts | Critical risk judgment, incident investigation leadership, safety culture | Experience → Transformation |
| Jane Carr — HR/IR | Policy drafting, award interpretation, workforce analytics | Union negotiation, enterprise bargaining, employee relations judgment | Experience |
| Kevin Arnold — Industrial | Tender pricing models, schedule templates, procurement specs | D&C/EPC delivery leadership, client trust, field team management | Experience |
| Stewart Furness — Services | Network build scheduling, subcontractor performance dashboards | Telco client relationships, integration leadership, field operations | Experience |
Developed by Walter Adamson (OutcomesNow). This dual-axis framework separates the business case for AI automation (IMPACT) from the technical feasibility of generative AI execution (FAVES), producing a 2×2 decision matrix that is substantially more actionable than single-composite scoring approaches.
Inspired by the Trust Insights TRIPS framework (Christopher S. Penn & Katie Robbert), which pioneered task-level AI prioritisation scoring. IMPACT × FAVES extends this foundation by identifying that most TRIPS dimensions measure economic opportunity rather than technical feasibility, and introducing a dedicated feasibility axis (FAVES) with factors absent from prior frameworks: output verifiability, systems-of-record entanglement, consequence asymmetry, and step decomposability.
| Dimension | What It Measures | High Score Means |
|---|---|---|
| Investment of time | Hours consumed per task occurrence | Large time savings opportunity |
| Multiplicity | Frequency × headcount | Compounding cumulative savings |
| Pain | Tedium, error-proneness | Faster voluntary adoption |
| Accrued economic value | Revenue or cost impact | Stronger ROI case |
| Cost of current execution | Fully loaded cost | Clearer payback |
| Trend | Growing, stable, or shrinking volume | Compounding future savings |
| Dimension | What It Measures | High Score Means |
|---|---|---|
| Fidelity of verification | Can a non-expert confirm output correctness? | Hallucinations detectable — safer to automate |
| Autonomy of context | Does all required context exist in accessible digital data? | AI has everything it needs |
| Volatility of consequence | How bad is a wrong answer? (inverse) | Room for AI error without damage |
| Entanglement with systems | System-of-record dependencies (inverse) | Standalone — no integration barrier |
| Step decomposability | Can the task be broken into verifiable steps? | Errors caught before cascading |
Role resolution: O*NET-SOC taxonomy (923 data-level occupations). Task inference: Company website, LinkedIn executive profiles, sector knowledge (Australian power/comms infrastructure), O*NET task lists for matched SOC codes. AI benchmark calibration: Anthropic Economic Index (January 2026), METR evaluations, frontier model benchmarks as of March 2026. Australian adjustments: Award wage complexity (IMPACT uplift), professional licensing regimes (FAVES V-score depression), geographic adoption lag (~6-12 months behind US market).
This analysis covers executive-level roles only. The majority of GenusPlus's 1,178 employees are field operatives, linespeople, electricians, and project engineers whose AI exposure profile is fundamentally different (much higher Atoms constraint). Task lists are inferred from public information and may not reflect the actual day-to-day priorities of each executive. FAVES scores represent a point-in-time assessment — AI capabilities are improving rapidly and scores should be recalibrated at least annually. The analysis uses the IMPACT × FAVES dual-axis framework which, by design, separates business case from feasibility — comparisons with single-composite frameworks will yield different rankings.