The Old Scenario Playbook Is Dead. Here's the AI Upgrade – Part 2: Three Techniques That Will Transform Your Scenario Sessions

Having a two-part blog is interesting.

Since Part 1 dropped, I've had lots of messages from readers who've tried the basic framework. One advisor ran scenarios for their corporate strategy in 60 minutes that would have taken their consultants three months to deliver. Another reader tried it and felt that results were a bit... meh. Mileage may vary.

This is an important point about LLMs. Michaelangelo didn't just swing a hammer at a block of marble and reveal David (hey, if you're going to use a metaphor, why not use one with lofty ambition?). The prompt frameworks I share in these blogs are that marble block - it's a solid starting shape. Iteration is the chiseling. Test a prompt for your use case. Evaluate the output, then refine (e.g. add constraints, examples, domain specifics). Non-determinism is the "stone's surprise". One minute you might get an unusual split here for a more poetic answer another you will get a fracture there where the LLM misses the point.

In the context of last week's blog for example, with one scenario I got much better results when I asked the LLM to channel a particular local expert and author in its response.

Hopefully the tortured metaphor conveys the point.

On to this week. One of the most interesting messages came from one risk manager: "The scenarios are great, but how do I make them stick? Typically when I've done this before, people's eyes glaze over. They need to feel it, not just read it."

Exactly.

Generating scenarios is just the beginning. If you can make those futures visceral and actionable they stand a much better chance of engaging your audience. In this Part 2 I'm going to share three prompting frameworks that transform scenario planning from an intellectual exercise into something that turbo-charges how organisations engage in them.

With chisel in hand, let's dive in.

Technique 1: Backcasting - Engineering Tomorrow by Working Backwards

Most scenario planning looks forward and asks "what might happen?" Backcasting flips this entirely: define your optimal future state, then work backwards to map exactly how to get there.

This is the generalisation of the "reverse stress testing" that risk managers have been using for years - "How can I lose $100m?".

Here's the prompt framework that is very.... Marty McFly:

Making Backcasting Actionable

Next, you stress-test your backcast. After generating your reverse timeline, run this follow-up:

When I ran this process for the paid research scenario from Part 1 of this blog, the AI identified that the strategy pivot required hinged on network effects. What wasn't obvious looking forward, but when working backwards, it became clear that without reaching a threshold number of platform contributors by mid-2026, the entire timeline would collapse.

Result? The implementation plan can be revised to address the key fragilities in the strategy.

The Pre-Mortem Twist

The final step in our backcasting framework boosts results further. Once you have your reverse timeline, run a "pre-mortem from the future":

This combination of backcasting, testing key assumptions plus pre-mortem produces some very rapid insights into the strategic choices (or risk scenarios for a big chunk of my readership) you are facing.

Again, through this process, iterate ruthlessly, reference key experts, authors or books that you think will better reflect the specialist knowledge for your domain to maximise the insights this approach provides.

Technique 2: Living Artefacts - Making Tomorrow Tangible Today

Abstract scenarios bounce off executives' mental armour. But show them tomorrow's Financial Times headline about their company's collapse? Now you have their undivided attention.

This isn't about pretty pictures. It's about using generative AI to create visceral experiences of possible futures that bypass intellectual defences and trigger genuine strategic thinking. I've used media "injects" in scenario workshops before and they are very powerful but they can be very time consuming and expensive to produce. Result? We tend not to use them very often. And some of the video content now possible simply wasn't available before the current AI capability (the newscast clip I made to post on LinkedIn for this blog being a prime example).

The Artifact Toolkit

Start with this media generation framework:

I did this with a cyber attack scenario for a major global bank. It laid out some hard hitting artefacts which can then be turned into images, news reports or other artefacts with your favourite LLM tool for maximum realism. Once it had finished, Chat GPT asked: "Would you like me to extend this into a “day 2” narrative (e.g. analyst notes, client lawsuits, board resignations) or keep this focused on the immediate unfolding crisis?" It read my mind because that's exactly what the next part of the technique explores.

The Emotional Gradient Technique

Don't just create one version. Generate an emotional spectrum:

Again with the cyber scenario, ChatGPT came back with "Do you want me to simulate internal board minutes for each scenario (showing how directors framed their choices and trade-offs) so you can see the anatomy of “good, mediocre, and bad” decision-making side by side?".

That's the power of making futures tangible.

Dynamic Scenario Immersion

For maximum impact, and to prepare for timed "scenario injects" through the course of a workshop simulation, create a "day in the life" experience:

I added in a "Think deeply before replying" at the end as in this session I felt as though ChatGPT was getting a little trigger happy. This produced a detailed response for each of the emotional gradients earlier. Nice.

For each of these take the initial prompt and use your favourite tool (Google's Nano-Banana or Veo3 for example) to create a rich set of media to use as injects in your scenario workshop. Here are some quick zero-shot examples from the paid research scenario again:

These media can be used, as a scenario unfolds during a workshop, to add realism to the deliberations of decision makers in the process.

Technique 3: Weak Signal Amplification - Finding Tomorrow's Disruption in Today's Noise

Our third technique focuses on the fact that every disruption starts as a weak signal. Kodak ignored digital cameras. Blockbuster dismissed streaming. The signals were there - they just seemed insignificant.

AI can detect patterns in weak signals that human analysts might miss, then amplify them into full scenarios before they become obvious to competitors.

The Signal Scanning Protocol

Start by casting the widest possible net:

I recommend running this with the LLM's "deep research" setting or equivalent for the best results and be very clear on the industry / market segment that you want the LLM to focus on. I did this for the Governance, Risk and Control (GRC) Technology sector, an area I know well, and the results were impressive, giving food for thought.

The Amplification Engine

Once you've identified weak signals, amplify the most intriguing ones to go deeper:

Here I took the Gen-Z communication preferences from the weak signals and fed it into the amplification engine for some fascinating potential trends in the GRC Tech sector!

The Collision Matrix

Further deep insights come from colliding weak signals:

Putting It All Together: The Strategic Synthesis

A final step threads the outputs together. Again, inject your own expertise, thoughts and experts to channel in this final prompt. Run this synthesis to connect everything:

The Bottom Line: From Passive Prediction to Active Creation

These techniques aren't just designed to help you see the future - they help you create it. When you combine backcasting's reverse engineering, artefacts' emotional impact, and weak signal amplification, you're not just planning for scenarios. You're elevating your organisation's decision making. Whether that's a market expansion or thinking through a supply chain disruption scenario

Among the many ways that AI can augment current capabilities, I firmly believe that the companies that master these techniques won't just survive disruption. They'll be the ones doing the disrupting. The ability to think through risks and strategies quickly or produce engaging workshop materials allow us to get our best thinking applied to uncertain futures. AI allows us to embed this as regular practice and not just an occasional luxury.

Until next time, you'll find me at dawn, amplifying weak signals and discovering that the future is already here - we just haven't been looking in the right places...

Resources and Further Reading

Historical Context:

• Kahn, H. (1984). Thinking About the Unthinkable in the 1980s. Simon & Schuster.

• Schwartz, P. (1991). The Art of the Long View: Planning for the Future in an Uncertain World. Doubleday.

• Shell International. (c. 1970s). Internal Scenario Planning Documents. Shell Archives. (Note: Specific documents are accessible through the corporate archives).

  
  

Modern Application:

• Govindarajan, V. (2016). The Three-Box Solution: A Strategy for Leading Innovation. Harvard Business Review Press.

• Ramirez, R., & Wilkinson, A. (2016). Strategic Reframing: The Oxford Scenario Planning Approach. Oxford University Press.

AI Integration:

• Anthropic. (n.d.). Guide to Using Claude for Strategic Planning. Anthropic. (Note: This refers to documentation and articles available on Anthropic's official website).

• Boston Consulting Group (BCG) Henderson Institute. (2024). Augmented Strategic Planning. BCG Publications.