Applied AI field guide 002: The gameplan, AI innovation

Everyone has a plan until reality hits.

Field guide 002 is the gameplan BTS uses with clients scaling AI innovation. It holds one objective: a human driven, self-reliant, constantly renewing applied AI culture and skillset. Not a robot takeover. Not a technology installation. Not something handed to a consultant.

Flywheels and diamonds

A flywheel is a small team using AI to build one specific innovation. It goes narrow, goes fast, gets results, and communicates what it learned. A diamond is the flywheel with organization-wide impact, and a diamond gets mandated. The guide walks the full arc from first flywheel to institutionalized new way of working.

What's inside

• Three pillars, with real use cases. Individual productivity, team workflows, and AI-native products and services. Each pillar comes with what AI does, how operations improve, and specific places to start.
• Where flywheels come from. Your AI superusers, your enthusiasts, your volunteers, and your shadow AI builders. Consider an AI amnesty month.
• What to skip. Readiness and gap assessments, 2x2 prioritization matrices, and multi-year plans. The AI innovation cycle runs about seven months.
• The chain reaction. Five moves to spread innovation beyond pockets: coach the coaches, run vibe-a-thons at 1,000 people or more, peer-to-peer showcases, deliberate FOMO, and the ten sparks.
• The diamond checklist. Eight tests an innovation passes before you declare it, including the SpaceX design algorithm so you avoid automating steps you should be deleting.
• Ten steps to institutionalize a diamond. Peer demos, workflow simulations, new metrics, new team structures, a cutover date, and the hard conversations arriving with them.
• Tokenomics and renewal. Hire agents, measure them, retire them when they stop earning their keep.
• Lead like a gardener. Eyes on, hands off. You create the conditions. You do not make the plants grow.

What you'll do differently

• Launch flywheels in weeks, with no use-case roadmap and no prioritization exercise first.
• Tell the difference between a nice pilot and a diamond worth mandating.
• Give your core team one job with three parts: build, cheer, equip.
• Retire innovations on a schedule as better capability arrives.