I’ve observed that many people and organizations are confused—if not completely hypnotized—by the relentless pace of AI disruption. Some fail to appreciate the magnitude of the shift. Others understand the gravity but are frozen by analysis paralysis. And the few taking action often react defensively rather than proactively ("I guess we have to deploy Copilot to all our employees now, right?").
Since people trust me to mentor their teams, I feel a strong responsibility to prepare them not just for today’s challenges, but for tomorrow’s ambiguity.
The core problem is that organizations try to treat AI adoption as a linear plan. It isn’t. It is a complex adaptive challenge.
Because we cannot predict exactly how AI will evolve over even the next few months, and how people (stakeholders, employees, customers, ...) will respond, we cannot "plan" our way to success in the traditional sense. We must embrace uncertainty. We need short feedback loops, safe-to-fail experiments, and rapid learning.
I designed this specific workshop to help teams do exactly that. It acknowledges that while no one "knows" the future, we can still take meaningful actions to maximize value and minimize risk.
Here is how to run it.
Step 0: Framing
Define a timebox and select your audience. I recommend taking about 2 hours at minimum, and 4 to 8 hours ideally. How much time you need for a productive discussion that doesn’t feel rushed will depend on the number of participants, the diversity of perspectives, to what degree the participants are already aligned, etc.
The workshop works for all sorts of teams: individual development teams, executive teams, and any community of people sharing similar challenges or goals.
Step 1: Setting the Stage
Gather the participants and establish the "Time Horizon." Are you discussing the next few months, or the next 2 years?
- High Urgency: If you are under pressure from competitors or market shifts, favor shorter horizons (1-6 months).
- Strategic Focus: If the room is full of executives setting long-term direction, favor longer horizons (1-2 years).
Note: I do not recommend looking beyond 2-3 years. These days, anything beyond that is likely a waste of valuable discussion time.
Step 2: Gathering Insights
Give participants quiet time to write down events, changes, or milestones they believe might occur as a result of the AI explosion within your chosen Time Horizon. Encourage them to think broadly—technical, operational, legal, and cultural.
Examples:
- “Our Level 1 Customer Support becomes fully automated.”
- “AI coding agents allow us to increase feature development throughput by 500%.”
- “One of our agents causes significant user data leakage.”
Step 3: Building Shared Understanding (The Matrix)
Draw a large 2-axis diagram visible to everyone.
- Horizontal Axis (Likelihood): From "Highly Speculative (<10%)" on the left to "Almost Guaranteed (>90%)" on the right.
- Vertical Axis (Desirability): From "Massive Downsides" at the bottom to "Massive Upsides" at the top.
Have participants plot their items one by one. This is where the magic happens. Allow for debate, but don’t get bogged down in precision—we are mapping uncertainty, not doing accounting. If the group radically disagrees on the position of an item (e.g., "Automating Level 1 support"), considering breaking it down. For example:
- "Cost of Level 1 support drops to nearly zero" (High Desirability)
- "Need to invest in re-training our Level 1 Support staff" (Low Desirability)
Step 4: Exploring "What to Do"
Now that you have mapped the landscape, how do you act? Use the quadrants to define your strategy.
1. Top-Right (Likely & Beneficial): EXPLOIT
Key Question: "How do we reap these benefits immediately?"
2. Top-Left (Unlikely & Beneficial): PROBE
Key Question: "How can we increase the odds of this happening?"
3. Bottom-Left (Unlikely & Harmful): MONITOR
Key Question: "How do we ensure this stays unlikely?"
4. Bottom-Right (Likely & Harmful): MITIGATE
Key Question: "How do we dampen the impact or buy time?"
What comes out of these discussions will greatly vary from organization to organization, from team to team. But here are a couple of examples:
“AI coding agents allow us to increase feature development throughput by 500%.”
Let’s say you have placed this item in the Top-Right quadrant. Participants may surface the following actions:
Allocate budget and empower all teams to experiment with any tool they want, as long as it meets compliance checklist
Invest in deployment pipeline automation as we don’t want deployment to become our new bottleneck
Pivot our hiring strategy to focus on developers willing and able to go beyond coding and into customer interactions, product design, product usage data analysis, etc.
“One of our agents causes significant user data leakage”
Let’s say you have placed this item in the Bottom-Right corner. Participants may surface the following actions:
Adopt strict “humans-in-the-loop” policy for every product change involving user data
Re-architect the deployment processes in order to enable canary-releases and rapid rollback procedures
Carry out an audit of our data encryption protocols
Step 5: Deciding What’s Next
By now, you will have a wall full of ideas and a shared sense of urgency. Do not create a 12-month roadmap. Instead, apply agile principles. Identify the most critical items and turn them into actions for your next Sprint or Quarter. Treat these actions as experiments: “We bet that doing X will result in Y. We will review the results in Z weeks.”
Final Thoughts
Given the pace of innovation in the AI field, and how emergent and unpredictable some of its applications are, you will get things wrong. If I had done this workshop with my teams the day prior to the release of Gemini 3, and done it again a few days later, I’m sure we would have moved a number of items from left to right, and several of our earlier conclusions would have become outdated as a result. That is the nature of complexity.
Do not get attached to the snapshot you create in this workshop. The goal isn’t to create a perfect plan, but to build the organizational muscle to sense change, and respond to it. Revisit this board regularly.
If you need help facilitating this conversation, or if you want an external perspective to help your organization navigate what’s to come, you know where to find me.