Field guide 003: Ten leadership sparks to accelerate AI innovation

Plenty has been written about why AI innovations fail. This guide is about why they succeed.

Most successes share one thing: The leaders ran a gameplan and used 10 deliberate actions to create the conditions for it. Call them levers, nudges, or pushes. BTS calls them sparks. Field guide 003 covers all 10, grouped into build, cheer, and equip.

The ten sparks

• Build. Coach, workshop, simulate.
• Cheer. Inspire, applaud, accelerate.
• Equip. Staff, tool, budget, secure.

What's inside

• The mirror test. Three statements to say out loud before you lead anyone else, and a reaction to date and revisit as your journey progresses.
• Acceleration benchmarks. Website prototype: one month to four hours. Complex financial model: one week to three hours. Synthesizing 500 feedback forms: two days to one hour. QBR prep: 40 hours to three. Use these to reset what you ask for and when.
• Nine simulations worth running, from P&L impact and competitor moves to talent model shifts and knowing when to fire an AI agent.
• Reliable muses versus hyped sources. A side-by-side for judging the AI success stories landing in your inbox. Named companies and narrow, detailed cases beat headline percentages.
• Four teams, one core team. Who staffs the core team, and why clearing barriers matters more than approving work.
• A four-step financial journey, including the trap of claiming hundreds of 0.25 FTEs in productivity and never seeing any of it reach the P&L.
• Why a data readiness project is the wrong first step, and what to build with the data you already have.
• Three spark-o-meters to diagnose a stalled program and dial the right sparks up.

What you'll do differently

• Ask for the prototype tomorrow instead of next month.
• Put on-demand coaching in front of scheduled training.
• Recognize innovators publicly, including the ones who took a smart risk and came up short.
• Fund AI innovation without freezing hiring across the board.