Framework
AI Strategy Stress Test
Run your strategy through the 7 Powers with AI in mind: traffic-light each power, then map what could help or hurt you over the next five to seven years.
Enormous money is flowing into "AI for X" companies promising to change how whole industries work. Whatever you're building, AI is now part of your strategy whether you chose it or not — it makes you operationally faster and cheaper to scale, but it also erodes moats you may be counting on. A complicated tech setup used to be a barrier; increasingly, it just isn't.
Step 1 — Traffic-light your 7 Powers
Take each of Hamilton Helmer's seven powers and ask how AI changes it for you against named competitors. Green: this power holds or strengthens under AI. Amber: it holds for now, but the barrier is thinning. Red: AI has eroded or will erode it — stop counting on it.
| Power | The AI-era question |
|---|---|
| Scale Economies | Does volume still lower your unit costs faster than a lean AI-native rival can drop theirs? |
| Network Economies | Does more usage still compound value, or can AI synthesise what your network used to be needed for? |
| Counter-Positioning | Could an AI-native entrant run a model you can't copy without breaking your existing business? |
| Switching Costs | How much of your lock-in survives when AI makes migration, integration and retraining nearly free? |
| Branding | Trust and preference may matter more in a world of infinite AI-generated alternatives — is yours real? |
| Cornered Resource | Proprietary data, distribution and relationships get more valuable; scarce technical talent gets less scarce. Which do you hold? |
| Process Power | Is your embedded way of working still hard to replicate, or is it exactly the kind of process AI commoditises? |
Step 2 — Map threats and opportunities, five to seven years out
Really imagine your market in five and seven years. Where is AI helping you — and where is it hurting? Plot the biggest of each on the Opportunity & Threat Map so the strategy conversation ranks them instead of listing them. A useful calibration: the panic that "AI will eat SaaS" turned out to be overdone — even the top AI labs still buy SaaS rather than build it — but the moat that made SaaS defensible has genuinely thinned. Neither doom nor denial; look power by power.
The output isn't a score — it's a decision. Which one or two powers will you deliberately build because AI strengthens them, and which will you stop leaning on because it doesn't?
Source / credit
Related capabilities
Related tools
Work with Ben
Want help installing this?
Outstride OS is the system behind Ben's founder coaching — pre-seed to Series C. If this page names something you are living right now, start a conversation.