Coming next: The Second Factory →

Back to Library
Adoption & Product⚠️ Pitfall

Licenses and Lip Service (When Everyone Logs In and Nobody Changes)

The Problem

You bought AI licenses for the whole company and built a course people can take at their own pace. Three weeks later the dashboard is a thing of beauty: 94% logged in, 88% finished the training. The CEO forwards it with three exclamation points. Then you ask people one question in the lunch line. What did you use it for this week? An email to a vendor. A summary of a long thread. A chicken recipe. "I haven't really had time yet. I did the training, though." Logins and completions are true numbers. They just don't tell you whether anyone works differently. Somebody played the whole course at double speed in another tab and picked the longest answer on every quiz. Indifference looks exactly like compliance, and it never shows up in the logs.

The Principle

Usage is the receipt. Logins, licenses and course completions are promises. Daily use on real work is proof. Rolling out AI looks like a technical problem: buy the tools, run the course, watch the numbers climb. Ronald Heifetz would call it an adaptive challenge. People have to change their habits, and nobody can install a habit for them. So measure the receipt gap: the share of people with access, minus the share using it every working day. Then close it the slow way. Sit beside people while they work on a real problem of their own. Get someone people watch to stand up in the front row. And never put the login number on a board slide without the daily-use number next to it.

🎯

4 Steps to Take Action

  1. 1

    Take the receipt gap this week

    Access minus daily use. An active day is one real interaction, not a login. A daily user is active every working day. Put both numbers side by side wherever the login number shows up today.

  2. 2

    Ask the lunch-line question

    Ask twenty people at random: what did you use it for this week? Count how many name a real piece of their own work. That answer shows you the indifference the logs miss.

  3. 3

    Trade a module for a sit-down

    Run a three-hour Bring Your Own Problem session. Everyone brings one real problem from their job and works on it in the room, with helpers who never grab the keyboard. Follow up with every person within 48 hours.

  4. 4

    Put someone visible in the front row

    Don't fill the first room with enthusiasts. Invite a respected skeptic, someone people watch. When they stand up, the room follows.

💡 Inspired By

Chapter 2 of The Second Factory, "Licenses and Lip Service." And a discovery from years ago at a large financial services company: about three percent of the people with access to an expensive, beautifully governed data warehouse actually used it. That's where "usage is the receipt" came from.

This is what Analytics Pitfalls & Principles is about: people helping people.

In Practice

At ClassicMotors, a 900-person classic car marketplace, the CEO stood up at the all-hands and announced the company would be AI native. Licenses for everyone on Monday. A six-module course. A hundred percent completion in thirty days. Behind him, a dashboard with two numbers: licenses activated, training completed. Three weeks in, the numbers were 94% and 88%. He was putting them on the first slide for the board. Nora Berg, the senior manager he'd put in charge, asked the data team for one more number. Not logins. How many people used the tools every working day? Eleven percent. And 34 of those had simply scheduled an inbox summary for six every morning, which most of them never opened. Module four of the course was called Building Agentic Workflows with Skills, Tools and Retrieval-Augmented Generation. A campaign manager finished all six modules in one afternoon and told her he still didn't know what he'd use it for. The numbers on the CEO's slide were true. They just didn't tell him anything. Nora told him: "I watched you fall in love with a demo once. I'm not going to let you fall in love with a dashboard."

Take it to work

Give this card to your AI

A skill is a file your AI keeps. Add this one and it knows Licenses and Lip Service whenever your work runs into it.

Then ask it: “Does my plan fall into Licenses and Lip Service?”

How to add it
  • Claude (a paid plan): Customize, Skills, +, Upload a skill. Pick the file you downloaded.
  • ChatGPT or Gemini: where your plan has Skills, upload the same file.
  • Microsoft 365 Copilot Cowork: Plugins, Add, Upload, then pick the Cowork file.
  • An AI at work that won’t take a skill? Open this card as text and paste it in.

All the AI skills, including ones for the moment before an AI, data or BI strategy, a dashboard or a readout goes out.

More on Adoption & Product

From the next book

In The Second Factory, a CEO is about to open his board meeting with 94% of the company logged in. The manager he put in charge of AI asked the data team a different question first.

This card is drawn from The Second Factory, Brian Cariveau’s next novel, about a company learning to work with AI. The opening scene is free to read now.

Heard something like this at work? See the pitfalls behind what else you’ve heard this month.Heard at work: free, 1 minute, no account