Start with the decision, not the technology.

“We know AI matters, but we do not know where to start or how ready we are.”

• —Executive and functional interviews
• —Review of strategy, process, data and technology documentation
• —Scoring across seven readiness dimensions
• —Opportunity identification and prioritisation workshop

• —Readiness maturity view
• —Opportunity heatmap
• —90-day priorities
• —Executive readout

Which opportunities to pursue first, and which capability gaps to close before investing further.

“Our AI projects are disconnected and do not add up to a strategy.”

• —Outcome and value-chain mapping
• —Decision, process and knowledge analysis
• —Digital work and automation potential assessment
• —Investment modelling and roadmap sequencing

• —Enterprise intelligence blueprint
• —Investment case
• —Staged roadmap

Where intelligence should be applied across the enterprise, in what order, and with what investment.

“Agents and models are being used without clear accountability or controls.”

• —Current-state governance and risk review
• —Accountability and decision-rights design
• —Agent and model oversight, approval and intervention design
• —Control and audit evidence framework

• —Governance framework and policies
• —Operating model and RACI
• —Control catalogue
• —Audit evidence design

How AI will be owned, approved, supervised and evidenced — and what must change before scaling.

“Our model choices and API costs are difficult to manage.”

• —Workload inventory and classification
• —Provider and model suitability assessment
• —Quality, latency, resilience and sovereignty analysis
• —Cost-per-task economic modelling

• —Target model architecture
• —Evaluation strategy
• —Routing and fallback principles
• —Provider recommendations

Which models and providers to use for which workloads, and how to evaluate and route between them.

“We are unsure what compute to provision, where, and at what cost.”

• —Workload demand modelling
• —Placement options: owned, rented, hosted, cloud, local and edge
• —Power, connectivity and resilience assessment
• —Scenario costing and utilisation modelling

• —Infrastructure scenarios
• —Demand-to-compute roadmap
• —Cost and utilisation model

What compute to commit to, where to place workloads, and when to scale.

“We cannot see or govern AI across departments.”

• —AI estate inventory across Mibono and third-party systems
• —Integration capability assessment per source
• —Common metrics definition: cost, value, quality, risk
• —Phased deployment design

• —AI estate inventory
• —Integration coverage map
• —Common metrics framework
• —Phased MIP deployment plan

Whether and how to deploy MIP, what it will cover at each phase, and what it will not.

“We cannot prove AI returns.”

• —Initiative and expenditure inventory
• —Adoption and usage analysis
• —Realised-outcome measurement against baselines
• —Portfolio rationalisation

• —Portfolio value review
• —Scale / fix / consolidate / replace / stop recommendations
• —Value measurement framework

Which initiatives to scale, fix, consolidate, replace or stop.

“We want agents, but need a sound plan.”

• —Task-level work analysis
• —Agent and digital worker suitability scoring
• —Human handoff and control design
• —Proof-of-value scoping

• —Digital workforce opportunity map
• —Handoff and control design
• —Bounded first proof of value

Which work to give to agents first, under what controls, and how success will be measured.

“We need an enterprise-wide strategy and architecture.”

• —Integrated strategy and value work
• —Work, model and compute architecture
• —Governance and operating model design
• —Implementation prioritisation with the executive team

• —Integrated enterprise intelligence blueprint
• —Target architecture
• —Governance model
• —Implementation roadmap and investment case

The full enterprise programme: what to build, buy, retain, govern and fund.