How do you run a scenario planning session?

This is foundational. Being able to have clear comms around what the subject and desired outcomes are. Include all stakeholders and get as crystal clear as possible. It’s always also best to document everything for reference.

I think it’s also helpful to scope what tools/resources will be available for the exercise. For example, AI is a good tool at every stage of these exercises in a variety of ways. At the outset it can help organize and summarize scope definition or help ideation around different scenarios. But if it’s not available to use in the org (ie ChatGPT banned and no in-house alternative) it is important to know that. In today’s day and age probably a good idea to have a sense of what resources should be in the go to toolkit for these exercises. Scope definition then clarifies which of the tools to pull out from the kit.

This is your brainstorm/brain dump. This is leaving no stone upturned. Analysis tools and frameworks are great but what’s most important is having the right data points to plug in to the model.

Completely agree with Cory’s comment above. And this ties in the toolkit from the first step. Example, determining you can extract data via AI that was previously not accessible could be a game changer and totally affect how you decide which drivers and factors to focus on.

Get creative. Let scenarios play out. Similar to user stories or archetype modeling, this is a place to really take things as far as you can go, and allow the forward thinking visionaries to align on best ways forward.

Something that is typically overlooked here is the thinking approaches that should be employed in creating the scenarios. Many scenarios end up with low forecasting value as they are victims of group think and bias. This risk is amplified if in the scoping step, recruitment of the participants was not carefully thought through. In any event it's always a good idea to decide what approach(es) to consider, e.g. Lateral Thinking, Counterintuitive Thinking, Divergent vs Convergent. Case in point, Post-Covid very few forecasters believed the world would experience a run of high inflation which would drive interest rates dramatically higher in a very short period of time. Scenario builders that assigned a higher probability to it had an edge.

Testing is an under-utilised art. Develop or utilise existing tools that allow your team to test potential scenarious rather than continuously debate their merits or weaknesses ad nauseum. This is a trap for governance boards: they do not ask leadership to develop evidence around a range of scenarios and then procrastinate on making decisions until a real problem is faced or everyone forgets about it until the next time it is raised.

AI is getting great at simulations and we’re not far off with quantum computing which is great at running simulations. Knowing how to leverage the available state of the art tools is a valuable leg up.

Storytelling is the key to communication. The prerequisite is that communication starts at the beginning and no one is taken by surprise. Keep the story and the medium in your hand. Live, with your voice, and create an arc from the beginning, which is very short, and continue to point out the touch points you had with the audience during the journey. The more vague you are, the more space the audience has to fill in their minds. So show a timeline and how the original goal was achieved. Avoid the nitty-gritty of deep details unless they are a good way of illustrating a general point. And stick to a limited number of scenarios to highlight. The story must have a satisfying ending, even if you are halfway through the roadmap.

Some thought should be given to how to incorporate scenario building into the organizational structure if it has not already been done. In an AI/Automation driven world, organizations that automate the process weaving together the different steps, data, decisions, monitoring, measurement and adjustment into an integrated system stand to gain an edge.

Don't forget to follow up. The world is full of one-off projects that had their 15 minutes of fame and then died because everyone moved on and stopped caring about pushing the issue, even though the challenge was still there. Ownership is only valued when the hot (or very cold) potato is being handled.

This is a very important point: to actually implement your scenarios and follow them up. Too many great scenarios never see the light of day - seen as too difficult to follow through on rather than unworkable.