Most agency AI workshops fail because they are presentations disguised as training. Someone shows slides about what AI can do, demos a few tools, and sends everyone back to their desks with a login and good intentions.
Two weeks later, nothing has changed.
An effective workshop is hands-on, role-specific, and produces output that gets used the next day. Here is how to run one.
Before the workshop
Choose one workflow per role. Do not try to cover everything. Pick the single highest-impact AI use case for each role in your agency:
- Account managers: meeting summaries and client research
- Copywriters: first-draft generation and content briefs
- Designers: brief structuring and image generation
- Developers: code generation and documentation
- Strategists: competitor analysis and data interpretation
If you need help identifying which workflow to tackle first, our AI agency audit guide walks through the process.
Prepare real work. Gather actual client work that needs doing this week. Not demo briefs. Not example projects. The workshop should produce output that goes into production.
Set up the tools in advance. Everyone should have accounts, logins, and access before the workshop starts. Nothing kills momentum like spending the first 30 minutes on IT issues.
The workshop structure (half day)
Hour 1: Context and mindset (45 minutes, 15 minute break)
Start by addressing the elephant in the room: AI is not here to replace anyone. It is here to handle the work nobody enjoys doing. Share the specific time savings you expect and what the team will do with that time (more strategic work, better output, or simply not working evenings).
Show one powerful example relevant to your agency. Not a generic demo. Take an actual piece of recent work and show how AI would have handled the time-consuming parts. When the team sees real work done in real time, the penny drops.
Hour 2: Hands-on by role (60 minutes)
Split into role-based groups. Each group works through their chosen use case with real work:
- The brief. Define exactly what the AI needs to produce and what inputs it needs.
- The prompt. Write it together, testing and refining as a group.
- The output. Generate it, review it, identify what needs human refinement.
- The workflow. Define when and how this fits into their existing process.
The goal is that each group leaves with a tested prompt and a clear workflow they can use tomorrow.
Hour 3: Share and commit (45 minutes)
Each group presents their workflow to the full team: what they built, what the output looks like, and how it saves time. This cross-pollination is valuable because it shows the breadth of applications and builds collective excitement.
End with commitments: each person states the one AI workflow they will use this week and what they expect the result to be.
After the workshop
Week 1: Daily check-ins. A quick Slack message each morning: “What did you use AI for yesterday? What worked? What did not?” This maintains momentum and surfaces problems early.
Week 2: Prompt library. Collect every prompt that worked well into a shared document. Our prompt engineering guide covers the framework for writing prompts that produce consistently good results. This becomes the agency’s AI playbook.
Week 3: Review session. 30 minutes as a team. What has stuck? What time savings are real? What needs adjusting? This is where you decide whether to expand to more use cases or refine the existing ones.
Week 4: Measure and decide. Compare the time spent on each workflow before and after AI. Share the numbers with the team. The data makes the case for continued adoption better than any mandate.
The one thing that matters most
The workshop is not the end. It is the beginning. The agencies that succeed are the ones that treat AI adoption as an ongoing practice, not a one-off event. The workshop creates the spark. The daily use, the prompt sharing, and the weekly refinement keep it burning.
This is part of Delivery Notes, a series on implementing AI inside your agency. Subscribe to the newsletter to get new articles weekly.
