Independent AI strategy and governance · Austria · EU
AI strategy your organisation can actually run.
I define the strategy, build the roadmap, and design the operating model and team. Then I prepare your Microsoft 365 estate, your AI agents and your evidence, so regulated mid-sized organisations across the EU can put AI into production and sustain it.
The gap
Most organisations use AI. Few can show it works.
of organisations say they are scaling AI across the enterprise.
McKinsey, State of AI 2026. Self-selected survey of 1,719 organisations.
report any impact on earnings.
McKinsey, State of AI 2026.
of 150 surveyed enterprises moved fewer than half their AI pilots into production last year.
Madrona, a venture investor.
The blockers named most often are integration, security and compliance review, and data access. Model quality is rarely the problem. That is the part I work on.
The shift that matters: from personal productivity AI to engineered, workflow-embedded AI.
- Used by individuals, on their own initiative.
- Quality depends on the person and the prompt.
- Value is real but hard to measure, and stays with the user.
- Runs on whatever data the user can reach.
- Built into a business process, with a named owner.
- Runs on governed data products and is tested against agreed measures.
- Guardrails, evaluation and an audit trail are part of the design.
- Value is measured in the process: time, cost, quality and risk.
The first is easy to start and hard to measure. The second is where lasting value sits, and it needs the strategy, data products, operating model and governance I work on.
What I do
Four ways to work with me
- DefineStrategyWhere AI creates value, and where it does not.
- PlanRoadmapSequenced waves with owners, costs and acceptance criteria.
- OrganiseOperating model and teamFunding, decision rights, roles and the mix of skills.
- PrepareReadiness and assuranceThe data, the estate, the agents and the evidence to carry it out.
- SustainGoverned deliveryOwners, reviews and evidence that keep it paying back.
AI strategy, roadmap and operating model
Define where AI should create value, turn it into a roadmap your teams can execute, and design the operating model, team structure and ownership that sustain it, including the move from personal productivity AI to AI built into your workflows. Fixed fee, handed over to named owners in your organisation.
AI readiness and managed governance
A fixed-fee assessment of identity, permissions, data maturity and architecture on Microsoft 365, in four to six weeks, with the data products AI depends on built where gaps appear. Then monthly governance that keeps agents inventoried, permissioned and evidenced.
Agent assurance
An independent evidence pack for each AI agent you build or buy: what it can reach, how it fails, how a person stops it, and what the audit trail shows.
Governance for your customers
Add governed AI delivery to your Microsoft practice without hiring a team. I deliver the assessment with your account lead; you keep the relationship.
How it works
One process, one price, and a hand-over to people in your organisation.
- 30 minutesCallYou tell me where the pilot is stuck. I say whether I can help. No charge.
- 3 working daysScopeOne written scope and one fixed price.
- Set in scopeDeliverThe strategy, roadmap and operating model, the readiness assessment (read-only access), or a data product build in an agreed environment.
- Final weekHand overThe roadmap and operating-model pack, or ranked findings, scripts and the evidence pack, with named owners and a rollout plan.
- OptionalSustainMonthly governance and refreshed evidence, or time-bound support until your own owners take over.
Who you work with
I have sat on the other side of the security review.
For more than 15 years I have built data, analytics and AI capabilities inside enterprises, from Red Bull to a health plan serving over a million participants. Most recently I deployed Copilot and data agents on Azure, and watched good pilots stall because identity, permissions and review evidence were not ready, not because the model failed. I have also led data and AI teams of up to 14 people, so the roadmaps, operating models and team structures I design are ones I know how to run.
Free download · English
The AI pilot-to-production readiness checklist
About 20 checks on two pages: identity and access, data exposure, agent inventory, review evidence and operations. Use it before your next security review.
Tell me where your AI work is stuck, or where it should go next.
In one 30-minute call I will tell you what it would take to get it into production, and to keep it there.