- Savings delivered
- €10M+
- Employees impacted
- 25K+
Selected work.
- Idea shortlist
- Fact-checked script
- Storyboard
- Assets
- Render and metadata
GenAI for online content creators
A pipeline that turns a creator’s idea into a finished video, with human-in-the-loop approvals.
- Depot gate
- On site
- In use
- Return gate
Equipment tracking with IoT devices
From a 20-line requirements list to an RFID business case with over €500K in identified annual savings.
- Workday
- Identity Manager
- Azure AD
- Access
Enterprise IAM, aligned onboarding and offboarding
A governed identity lifecycle for employees and external users across a 25,000-person organisation.
The model is the small part. The system around it is the work.
What I put into production
Typical shape
- Your data docs, ERP, tickets
- Retrieval search + permissions
- Model Claude or similar
- Guardrails policy, cost limits
- Human check where it matters
- In production owned, monitored
↺ Evaluation and cost tracking feed back into every step
Manual work saved
Share of the manual steps the system now handles, 12 weeks after go-live.
Cost per answer
Caching, routing and smaller models, month by month.
Where the effort goes
- Data & access 35%
- Integration 25%
- Evaluation 25%
- Model & prompts 15%
Share of a typical build. Model work is the smallest slice.
Twelve weeks from first call to a system your team owns.
I work embedded with your team. Each phase ends with something you can use, or a clear decision to stop.
A typical engagement
Weeks
€15–35K
A three-month project costs between €15,000 and €35,000, depending on the business scope and the architecture complexity.
Get in touchOther services I provide
AI implementation lead
I own delivery from pilot to production: scope, team, evaluation, rollout and the numbers that show it works.
Assistants, agents and pipelines
Hands-on builds on real data: retrieval assistants, agentic workflows and the plumbing around them.
AI audit and governance
A one-week diagnostic of what is running, what it costs and where the risk sits, with a plan you can act on.
Prefer to do it yourself?
The guides and templates I use with clients, packaged so your team can run them.
AI audit guideline
The checklist I use in a one-week AI audit: what to inventory, what to measure, where the risk usually hides, and how to write the findings up.
AI implementation template
A ready project plan from use case to production: phases, roles, evaluation set-up, go-live criteria and a handover pack.
Evaluation starter kit
Templates for building a test set from your own examples and tracking quality, latency and cost per answer every week.
Have an AI project that needs to ship?
A 30-minute call. No deck, no follow-up sequence, just whether I can help with what you are trying to do.