Top 3 AI Opportunities
The use cases with the greatest economic or operational impact. Prioritized by benefit, effort, and risk.
Practical Paper No. 01 · Workshop Guide
An AI Strategy in 120 Minutes.
For executive teams and boards of directors who want to finalize their AI roadmap in a single meeting—not in PowerPoint. Facilitation, hands-on workshop, clear results.
From Hype to Decision.
This paper distills what has proven effective in decision-maker workshops into a 120-minute format—without tool demos or technical debates. The result is a template for the next executive meeting, not just another set of strategy slides. Deliberately streamlined: one facilitator, one whiteboard, and one A4 template per use case. The discipline lies in the process, not in the materials.
Three results, binding.
At the end of the two hours, there must be three things on the table. If one is missing, the meeting is not over.
Risks & Dependencies
What needs to be clarified before a pilot: data, compliance, rights, supply chains, and personnel.
1–3 Next Steps
Include the owner, the date, and a specific first step. No empty phrases.
Participants.
Four to eight people. Fewer is okay, but no more—otherwise, the decision-making intensity drops measurably.
Executive Management / Division Management
Decides on the budget, resources, and responsibilities. Without this role, the workshop is just a casual chat over coffee.
Key Technical Expert
Has in-depth knowledge of the processes that AI is intended to transform. Often a division manager or senior operations manager.
IT / Data / Automation
Assesses the data available, interfaces, and realistic effort requirements. Prevents pipe dreams and calls for concrete architectures.
Legal / Compliance / HR (optional)
When personal data, regulated processes, or employees are involved: be there from the start, don't leave it for later.
Duration: 120 minutes.
Seven blocks. Each with a specific outcome.
TimeBlockResult
0–10 min.Vision & RulesA shared understanding of the purpose, scope, and decision-making framework.
10–30 min.AI Use Case ScreeningFive to ten potential use cases identified, divergent, unfiltered.
30–55 min.EvaluationTop 3 prioritized based on benefit, feasibility, and risk.
55–75 min.Strategy AlignmentAlignment with strategy, processes, data, and resources.
75–95 min.Risks & GovernanceRelevant risks, approvals, guidelines, and open questions have been clarified.
95–110 min.Action PlanFirst steps, owners, timing, and dependencies have been documented.
110–120 min.Decision & ConclusionBinding next steps confirmed; owner identified.
Five questions that guide every discussion.
If the discussion drifts into too much detail or veers into generalities, these questions will bring the group back on track.
- Where is the greatest economic or operational leverage?
- Where does work repeat itself often enough to make AI worthwhile?
- What data is available, maintained, and usable?
- What is realistic to achieve in a pilot program over three to six weeks, rather than over twelve months?
- What risks must be addressed before a pilot project begins?
Use Case Canvas: one page per use case.
The eight fields ensure that everyone on the team thinks at the same level of depth and prevent charismatic ideas from simply slipping through. If a field remains blank, the use case is not ready for a decision.
Problem / Opportunity
What problem are we solving, or what opportunity are we capitalizing on? A sentence without a solution.
Target Audience / Process
Who, specifically, benefits? What process is being changed?
Expected Benefits
Measurable or observable in CHF, hours, quality, or speed.
Data Availability
What data already exists? Where is it located, what is its quality, and how can it be accessed?
Risk / Compliance
Data Protection, Regulatory Issues, Rights, Reputational Risk.
Effort / Dependencies
Rough estimate in person-days or CHF, critical dependencies.
Owner
One person, specifically, not a department.
Next Step
A specific action within the next 14 days.
Facilitation: Three Principles.
The quality of the outcome depends almost entirely on how the discussion is moderated. Anyone who allows technical discussions to get bogged down in details will lose the decision.
First collect, then evaluate, then decide.
Do not mix these three modes. Evaluation kills collection; determination kills evaluation.
Effect before tool.
Which tool, which model, which provider: none of that matters in this round. What counts is impact, accountability, and feasibility.
Mandatory ownership.
Each action is assigned a specific person and a date. Without both, the action does not exist.
Sources & References.
- swidoc — Template for AI Workshops swidoc.ch
- marckarl — The 1-Hour Workshop on AI Prioritization marckarl.com
- BBV — AI Fundamentals Workshop for Decision-Makers bbv.ch
- Deloitte — C-Suite Guide to the Potential Value of AI deloitte.com
- OECD — AI Principles oecd.org
- AI Academy — AI Strategy Workshop for Decision-Makers, Zurich akademie-ki.ch
- MIT Sloan Executive Education — Developing an Effective AI Strategy executive.mit.edu
Guide (PDF).
Printable, A4. Bring it with you to your next meeting. You can find the fillable templates on the "Working Papers" page.
To the Working Papers