You’ve probably seen this before: your best salesperson quits, and everything they knew—the exact words that closed a deal, the follow-up rhythm that kept customers coming back—walks out the door with them. You stare at the renewal rate, but have no clue why that star performer made a client swipe three thousand bucks on the spot.
That’s the real pain for any brick-and-mortar store. It’s not rent or customer acquisition—it’s people. And the solution isn’t to monitor results harder. It’s to start capturing the process that leads to those results.
Let me introduce you to Jia Hua, founder of Aimega Pilates in Beijing. Four studios, fifteen employees, annual revenue of twelve million yuan. She did something that most peers haven’t even thought about: she stopped obsessing over outcome data and started systematically collecting process data.
Step one: put a recording device on your top performer. Jia Hua equipped her sales champion with a simple recorder, then fed the entire sales conversation into an AI tool. The AI extracted exactly what worked—specific phrases, question sequences, follow-up pacing. She got a dozen key actions out of it, then trained the whole team on them. Even if each person only nailed four or five, efficiency doubled. That’s how her renewal price per customer went from 23,000 to 30,000 yuan in a tough market. Not by pushing people harder, but by using AI to decode the champion’s method so everyone could hit 80 points.
Step two: let AI take over repetitive tasks, and process data accumulates automatically. Another pain point: coaches used to spend hours writing monthly training plans for clients, often staying up past midnight. And once written, those plans disappeared—no trace, no feedback loop. Jia Hua wrote a prompt template. Now coaches paste in the client’s info, and AI spits out a draft in five minutes. More importantly, every AI-generated plan is a piece of process data: which questions got asked, which got missed, where different coaches deviate. The curriculum optimization that took a year and a half before? AI finished it in one month.
Step three: make the tool so simple that anyone can use it. The key insight Jia Hua learned: personal use of AI and enterprise use of AI are completely different paths. For a team to adopt it, the barrier must be near zero. She built a dumbed-down sales simulation tool using Codex. Coaches can practice sales scenarios anytime, get AI feedback, no boss needed. The more they use it, the more process data piles up. AI gets smarter, and the cycle feeds itself.
Here’s the bottom line. Outcome data tells you what happened. Process data tells you why it happened. And only when you know the why can you know what to change.
Jia Hua’s parting words stick with me: “Do what AI can do well, then amplify the offline experience that AI can’t touch. Put them together, and that’s your competitive edge.” So next time you look at a spreadsheet full of numbers, ask yourself: where’s the process data? That’s where the real leverage lives.