On-site AI workshop: a strategic working day with your management team
When a management team workday is worth more than a 12-week consultancy engagement
Many leadership teams are stuck in the same discussion: we need to do something with AI, but what exactly? A twelve-week Big4 strategy engagement delivers a report, not a pilot. A hackathon delivers prototypes, but no prioritisation. An AI course for your frontline staff comes too early if the strategy hasn't been set. Our on-site workday is a structured strategy session with your management team, working towards an AI roadmap, a chosen pilot and a business case you can start acting on first thing on Monday.
Plan your workday View the day's scheduleWho this workday is for
This is not a training course for developers or a workshop for frontline staff. It is a working session for the people who decide on budget and strategy.
An AI strategy day works best with a compact management team of four to eight people who have direct decision-making authority. Think of the board, the COO, the CFO, the CIO or IT director, and business unit managers with P&L responsibility. An innovation lead or digital transformation manager is welcome to join, but the group needs to carry enough weight to give a go/no-go on a pilot at the end of the day.
What we explicitly do not do on this day: technical deep dives for your IT team, broad staff training or hands-on tool demonstrations. For that, we have a separate AI course for businesses in which your employees learn to work with ChatGPT, Copilot and related tools. The strategy session and that training can follow each other, but they are two separate products for different audiences.
We bring two people: an AI strategist who facilitates the discussion and an experienced technical lead who can assess feasibility and architecture. No team of consultants, no junior note-taker slowing down the conversation. A compact group, short feedback loops, a quick route to a decision.
What you take away at the end of the day
No hundred-page report that disappears into a drawer. Three concrete deliverables your management team can use straight away.
AI roadmap on A3
One clear A3 with all your candidate use cases, clustered by horizon (now, 6 months, 12-24 months). For each use case: a short scope, estimated value, data-readiness score, AI Act risk category and dependencies. Designed as a communication document for your board, for management team members who weren't there and, later, for your IT organisation.
Pilot choice with scope
One chosen pilot project, with an explicit statement of what is and isn't included. This covers success criteria, required data, integration points with existing systems and a rough estimate of lead time and effort. Not a vague intention, but a clearly defined assignment that can serve directly as an RFP or internal project proposal.
90-day business case
A well-founded business case for the first 90 days of the pilot: investment, expected savings or revenue, key risks, exit scenario and a three-year TCO indication. Not a spreadsheet with ten tabs, but a document that lets your CFO give a yes or no within a single meeting.
On top of that, the day also produces a number of soft outcomes that are often just as valuable: a shared vocabulary within your management team (what is an LLM, what is an agent, what is RAG), stated assumptions and risk tolerances, and clarity on who within the company will own the AI roadmap.
Schedule for the workday
Eight hours, four blocks, one deliverable per block. We work with use-case canvases, prioritisation matrices and a business case template that is filled in on the spot.
AI trends and use-case mapping for your business
A brief overview of generative AI, agents, RAG, fine-tuning and multimodal models, covering only what is relevant to your sector. We then map your key processes onto a use-case canvas: where manual work sits, where lead times are too long, and where customers drop out of the process. The aim of this session is a longlist of twenty to thirty candidate use cases.
Sector examples and analogous cases
We show you what comparable organisations in healthcare, retail, finance or industry have already rolled out, and the obstacles they ran into. These are not fictional cases but public examples from annual reports, conferences and open-source implementations. Your management team can see which patterns are relevant and which are not, and then extends or trims the longlist based on this input.
Prioritising on ROI and feasibility
The longlist goes through a two-dimensional prioritisation matrix: business value on one axis, technical feasibility on the other. For each candidate use case we score data readiness, AI Act risk category, vendor lock-in risk, GDPR/DPIA impact and organisational readiness. The result is a prioritised shortlist of three to five use cases, with explicit justification.
Pilot selection and 90-day business case
From the shortlist we choose one pilot. Together we complete a business case template covering investment, expected impact, success criteria, key risks, dependencies and a 90-day timeline. By the end of the session you have a document your CFO and board can say yes or no to, not in two weeks after a follow-up, but right then.
There is time for lunch and coffee between sessions. We work with whiteboards, sticky notes and a laptop, with no hour-long presentations. Your management team is actively involved all day, not sitting and listening.
Comparison with a Big Four AI strategy engagement
Same question, different approach. For some organisations, an engagement with a large consultancy is the right choice. For most SMEs and mid-sized organisations, a single working day is more effective.
| Aspect | Big Four AI strategy | Appfront on-site working day |
|---|---|---|
| Lead time | 8 to 12 weeks | 1 day, with about 2 weeks of preparation beforehand |
| Consultancy staffing | Partner, manager, 2-3 consultants | 1 strategist and 1 technical lead |
| Deliverable | Strategy report (50-100 pages) | A3 roadmap, pilot scope, business case |
| Implementation | Separate engagement, often with a different firm | We build the pilot ourselves, with no handover |
| Decision | After a presentation to the steering committee | At the end of the working day |
| Suited to | Listed companies, >5,000 employees | SMEs and mid-sized organisations, 50-1,000 employees |
We are not better than a Big Four firm, we are different. A twelve-week engagement adds value when multiple business units and hundreds of processes need to be analysed, or when compliance programmes require external validation. For a management team that wants to set an AI roadmap and a first pilot, a working day is faster, cheaper and more hands-on. And after the working day, a team is ready to actually build the pilot. See also our pages on AI development and enterprise AI implementation.
Sector-specific approach
Use-case mapping differs by sector. We tailor the sector examples, regulatory context and typical pitfalls to your industry.
Care and healthcare
Use cases around intake automation, record summarisation, scheduling and logistics optimisation. Key contextual factors: GDPR and DPIA processes for patient data, NEN 7510, BIG registration aspects for decision-support models, and the stricter AI Act risk categories that often apply. We also look at which use cases fall outside the primary care process and can therefore be delivered faster.
Retail and e-commerce
Use cases around product descriptions, demand forecasting, customer segmentation, dynamic pricing and intelligent customer service. Specific focus areas: integration with PIM and ERP systems, handling seasonal effects and cold-start situations, and the trade-off between quick, generic LLM solutions and longer custom projects. We also draw on our experience with e-commerce development.
Finance and insurance
Use cases around document processing, fraud detection, risk classification, customer communication and compliance monitoring. Key themes: model explainability that regulators accept, AI Act high-risk classification for credit scoring or claims assessment, and the choice between on-premises hosting and the cloud. Here, governance is almost always a separate workstream in the roadmap.
Industry and manufacturing
Use cases around predictive maintenance, quality control via computer vision, production planning optimisation and knowledge management (making technical documentation accessible via RAG). Focus areas: integration with OT systems and SCADA, handling sparse data in small production runs, and the choice between edge deployment and central models. We also look at how this fits with your IT modernisation.
AI Act, GDPR and governance on the workshop day
An AI roadmap without a governance component is not a roadmap. For every candidate use case, we explicitly weigh the legal and ethical context.
AI Act risk categories
The European AI Act distinguishes between prohibited, high-risk, limited-risk and minimal-risk AI applications. For each use case on your shortlist, we map which category applies and which obligations follow from it, such as documentation requirements, conformity assessment and human oversight for high-risk applications.
GDPR and DPIA
Use cases that process personal data will in many cases need a DPIA (Data Protection Impact Assessment). We flag which use cases trigger this and what a realistic DPIA lead time looks like. We are not a DPO, but we make sure you schedule the legal check in good time rather than just before delivery.
Vendor lock-in and TCO
Every AI stack involves trade-offs: a Microsoft Copilot stack is quick to roll out but hard to move away from, while your own open-source stack demands more engineering but gives you ownership. We discuss make-or-buy, model portability and exit scenarios, and include three-year TCO in the pilot's business case.
Data sovereignty and hosting
For some use cases, hosting within the EU or within the Netherlands is a prerequisite. We discuss which models are available for on-premises, EU cloud and US cloud, and which trade-offs in capability and cost follow from each. We do not argue for any single option, but we give you clarity on what the choices mean.
Our technical lead knows the relevant frameworks (AI Act, NIS2, GDPR, NEN 7510) and knows when you need external legal advice. We do not give legal advice; we make sure the roadmap does not stall on a surprise in month three.
Frequently asked questions
Ready for a workshop day your management team will get something out of?
One day, your location, your management team, our methodology. At the end: an AI roadmap, a pilot choice and a business case. No report, but a decision.
Schedule an exploratory call