The Moat Machine: I Wrote About Product Strategy in 2022. In 2026 I Built the AI Plugin That Runs It.

A developer built PM Superpowers, a Claude Code and Cursor plugin that runs product strategy frameworks interactively rather than generating documents. The tool, which includes a strategic-moat skill covering eight moat types and a six-step /strategy session grounded in Rumelt's Why/What/How, grew out of a 2022 UX Collective article on the same framework. The developer positions it as a strategy-thinking aid rather than a PRD generator, aiming to help teams answer questions like "what's our moat?" in minutes.

In December 2022 I published a UX Collective article on a systematic way to develop and socialize product strategy. Six steps. Rumelt’s strategy kernel. VRIO for pillars. A path from diagnosis to communication. It got 52 claps. I was happy about that. Then I went back to work and spent the next three years doing product strategy exactly the way I always had: a Google Doc, a lot of tabs, and somewhere between four and eight hours I didn’t really have. Not because I didn’t know the frameworks. I knew them well enough to write about them. The problem was never knowledge. It was execution. Most AI-for-PM tools stop at generating a PRD. They help you write faster, not think straighter. That sounds helpful until you’re in a leadership review and someone asks the question that turns a polished doc into wet paper: “What’s our moat?” If your answer is an adjective — sticky, delightful, AI-powered — you don’t have a strategy. You have a slide. The hard part of product strategy isn’t drafting. It’s the work that makes a draft survive that room: Done properly, this takes hours. Done quickly, it produces analysis that wouldn’t survive a single hard question. So most of the time, it doesn’t get done at all — and the roadmap keeps shipping features that don’t compound into advantage. There had to be a better way in 2026. The 2022 piece — The systematic approach to developing and socializing product strategy https://uxdesign.cc/product-strategy-framework-a4d49bf5b265 — was a field guide. Useful on paper. Still slow in practice. Knowing Rumelt’s Why / What / How doesn’t magically create an afternoon where you can run VRIO without interruptions. In 2026 I shipped what that article implied: an interactive AI workflow that runs the frameworks instead of summarizing them in another essay. Write-up: I wrote about product strategy in 6 steps in 2022. I just built an AI that runs it. https://aniganti.medium.com/i-wrote-about-product-strategy-in-6-steps-in-2022-i-just-built-an-ai-that-runs-it-4c6375527a5c Repo: aniganti/pm-superpowers https://github.com/aniganti/pm-superpowers PM Superpowers is a Claude Code / Cursor plugin: skills, one sub-agent, and a shared library of named strategy frameworks. It triggers in conversation — you don’t copy-paste a mega-prompt from a Notion doc you’ll forget to open. It is not a PRD generator. That distinction matters. Faster docs are easy. Defensible strategy is scarce. In practice it feels like a strategy-minded chief of staff who’s read VRIO, aggregation theory, and every pre-mortem post-mortem — and asks the next hard question instead of dumping a wall of text. The AI doesn’t freestyle a plausible-sounding essay; it interviews you along the framework’s actual dimensions. Brand hook, said plainly: answer “what’s our moat?” in minutes. The strategic-moat skill walks you through eight moat types, asks for evidence not vibes , rates strength, and surfaces deepening opportunities. You’re pushed to say what product behavior, data, or user dynamics support each rating — and where the moat is thin. You leave with a moat profile you can defend: strongest moats, weakest moats, deepening moves, and a defensibility verdict grounded in your product, not a generic essay about “network effects in tech.” That’s the moat machine. Not a slogan — a session you can run before the next strategy review. /strategy — the six-step session /strategy is the orchestrator from the 2022 framework, now interactive: Grounded in Rumelt’s Why / What / How. The AI interviews you; you don’t freestyle into a slide deck and call it strategy. Artifacts land in structured docs you can revisit — not chat scroll you’ll never find again. Standalone vrio-analysis applies the decision tree to concrete capabilities: Most teams skip a gate and still call it a pillar. The skill makes skipping awkward — which is the point. Moat analysis without competitive intel is fan fiction. The competitive-researcher sub-agent gathers competitive intelligence via web research so competitive-landscape isn’t “please paste everything you know about competitors.” You review the synthesis; you don’t become a tab farmer. That landscape then feeds VRIO and moat instead of living in a separate deck nobody opens. Recommended full strategy pipeline: