Scenario planning for technology uncertainty

0:00 Scenario planning

Here's a truth about forecasts: they're usually wrong in the details that matter. Scenario planning takes a different approach. Instead of asking what will happen, it asks what could plausibly happen, and how would we do in each case? In this lesson you'll learn the method, see a complete worked example for a marketing agency, and learn how to use AI as a sparring partner without letting it write generic futures for you.

0:32 What scenarios are

Scenario planning was popularised by Shell in the nineteen seventies, and governments and companies use it widely to prepare for uncertainty. Scenarios aren't predictions. They're structured, plausible stories that stretch your thinking, expose hidden assumptions, and help you find strategies that work across several futures, not just the one you'd bet on.

0:55 Why it matters

Why does this matter? Because the biggest strategic mistakes usually come from betting everything on one view of the future that turns out wrong. Technology change, regulation and customer behaviour are all uncertain right now, especially around AI. Scenario planning doesn't remove the uncertainty. It helps you find moves that work in several futures, hedges against the ones that would hurt, and early warning signs that tell you which future is arriving. That makes your organisation calmer and faster when things change.

1:31 Packing for uncertain weather

Here's an analogy. Scenario planning is like packing for a trip where you don't know the weather. You don't pack only for sunshine because the forecast says so, and you don't bring every piece of clothing you own. You think through a few plausible conditions, hot, cold, rainy, and pack items that work across them, like layers and a light rain jacket. Those layers are your robust moves. The umbrella you might need is your hedge.

2:04 Simple example: an online retailer

A simple example for a small online retailer. Focal question: how should we sell in 2029? Axis one: do customers mostly shop through AI assistants or still through websites and apps? Axis two: are delivery costs rising sharply or stable? Four worlds emerge quickly. And a robust move becomes obvious across all four: keep product data clean and structured, because both assistants and websites need it, whatever happens to delivery costs.

2:35 Seven steps

The method has seven steps. Set a focal question with a time horizon. List the driving forces: technological, economic, regulatory, social and environmental, using your horizon scan. Separate the fairly predictable forces from those that are both highly uncertain and highly impactful. Pick two of those critical uncertainties as the axes of a two-by-two matrix. Write four scenarios. Test your strategies against each. And monitor signposts.

3:04 Worked example setup

Let's do one. Focal question: what services should a thirty-person marketing agency in the UAE and Pakistan offer in 2030? Axis A: how much discovery happens inside AI assistants and agents, from a moderate share to a dominant one. Axis B: how strong trust and provenance regimes become, from weak, where synthetic content floods channels, to strong, where labelling and verification are widely enforced.

3:32 Four scenarios

That gives four worlds. Noise Wars: AI discovery dominates and trust is weak, so assistants gatekeep attention and brands fight for citations amid synthetic content. Verified Agents: AI discovery dominates and trust is strong, so assistants favour credentialed sources and agentic commerce grows. Content Glut: classic channels still matter but drown in cheap AI content. And Trusted Classics: search and social evolve slowly, and authenticity labels become a differentiator.

4:02 Strategy test

Now test strategies. Robust moves, good in every world: first-party data and owned channels, original research content, structured data and feed excellence, and AI-augmented production with human creative direction. Hedges, protecting against bad worlds: a provenance and disclosure compliance service, and brand-safety monitoring. Options, small bets that could scale: an agent-readiness audit product, and AI assistant visibility tracking. Tie early indicators, like the share of leads from AI assistants, to your horizon scan.

4:34 AI as sparring partner

AI is a great sparring partner here. It can suggest forces you missed, find inconsistencies in your scenarios, draft narratives from your bullet points, and role-play customers or competitors in each world. The lesson text has a stress-test prompt. But keep people in charge of picking the axes and judging plausibility, because models tend to produce generic futures that sound smart and fit no one.

5:02 Workshop and traps

Run it as a half-day workshop. Thirty minutes on the focal question and forces, using horizon-scan pre-work. Forty-five on ranking forces and choosing axes. An hour in small groups, each writing one scenario. Forty-five on the strategy stress test. And thirty on signposts and owners. Avoid the classic traps: best, worst and middle cases, which are just a forecast in disguise; axes that aren't independent; crowning one scenario as official; and never revisiting the work.

5:35 Make scenarios vivid

Give your scenarios vivid names and short stories. Noise Wars is memorable; scenario C is not. Write each as a short narrative from the point of view of a customer in that world, then describe what competitors are doing and what your organisation looks like. Vivid scenarios travel through an organisation and get used in everyday conversations, which is exactly what you want. Dry tables get filed away.

6:05 Signposts

Link each scenario to early indicators you can actually observe within a year. For our agency example: the share of client leads coming from AI assistants, how strictly platforms enforce AI labels, whether major platforms display Content Credentials, and regulatory action on AI answers. Add these to your horizon-scanning log. When indicators move, you'll know which world is emerging and can shift investment before competitors notice.

6:34 Three mistakes

Three common mistakes. First, choosing best case, worst case and middle case, which is just a forecast with error bars. Second, picking axes that move together, so your four worlds collapse into two. Third, holding a great workshop and never looking at the scenarios again. Tie each scenario's signposts to your horizon scan, and revisit the whole set at least once a year.

7:01 Try this now

Try this now. Write one focal question for your organisation with a time horizon, for example: what should our service mix be in 2030? List ten forces that could affect it. Mark each as fairly predictable or genuinely uncertain, and high or low impact. Pick the two most uncertain, high-impact forces that don't move together, and sketch a two-by-two with a name and three sentences for each world. Then test two of your current strategies against all four. Which ones hold up everywhere?

7:37 Recap

To recap. Scenarios are plausible stories, not predictions. Build a two-by-two from independent critical uncertainties, write vivid narratives with indicators, and test strategies for robust moves, hedges and options. Use AI to spar, not to decide. Your next step: run a mini scenario exercise on one focal question for your organisation. Next is the capstone: your three-horizon technology roadmap.