Field notesBook a call

Zipline/Field notes/

AI changes fast. People don't.

Every company has an absorption rate: the speed at which its people can take in a new way of working. Trust sets it, not technology.

Nick, ZiplineWritten by NickStrategy & Adoption Lead, Zipline

Every few months the AI models get noticeably better. Every few months, the distance between what they can do and what companies actually do with them gets a little wider. I have stopped seeing that as a technology problem. Every company has what I call an absorption rate: the speed at which its people can take in a new way of working. The model is ready on Monday. The organisation needs a year.

We have been here before

The economic historian Paul David told the best version of this story in 1990. Electric motors were available to factories from the 1880s, and for roughly four decades they made surprisingly little difference to productivity. Owners swapped the steam engine for one large electric motor and kept everything else as it was. The gains only arrived when a new generation redesigned the factory itself, with a small motor on every machine and a floor laid out around the flow of work.

Erik Brynjolfsson, Daniel Rock and Chad Syverson later gave the pattern a shape: the productivity J-curve. A general-purpose technology first demands investment in things no balance sheet records, such as new processes, new skills and new habits. Measured productivity dips before it climbs. Judge an AI project in the dip and you will cancel it.

Researchers have a name for the thing I am describing. Wesley Cohen and Daniel Levinthal introduced absorptive capacity in 1990 to explain why some firms profit from new knowledge while others, with access to exactly the same knowledge, do not. Their answer: a company can only absorb what connects to something it already understands. You cannot buy that. You build it by doing.

What resistance actually looks like

A 2026 survey by Writer and Workplace Intelligence found that 29 percent of employees admit to sabotaging their company's AI strategy. That number gets quoted as proof that people are the problem. I read it differently. It is what you get when a tool arrives without an answer to the question every employee is asking: what happens to me if this works?

The mirror image shows up in another number. In the 2024 Work Trend Index from Microsoft and LinkedIn, 78 percent of the people who used AI at work had brought their own tools, and more than half were reluctant to admit they used AI for their most important tasks. They are getting value from it and telling nobody, because being open might mean more work, a reprimand, or proof that their job can be done in half the time.

Amy Edmondson's research on psychological safety explains what is missing. Teams learn when it is safe to be seen trying, failing and asking. Someone who believes that showing a clever shortcut will cost a colleague their job will keep the shortcut to themselves.

Three promises

In our projects, absorption speeds up when the owner or director makes three promises, out loud and early.

  1. What happens to jobs. If nobody will lose their job to this, say it. If you cannot say it, people deserve to know that too, and you should not expect their help.
  2. What happens to the time. Hours saved are only good news if people know where they go: more customer visits, a new product line, going home on time.
  3. Who decides. People choose which part of their own work goes first, and a named colleague stays in charge of what the system produces.

There is good evidence for why the third promise matters. When Brynjolfsson, Danielle Li and Lindsey Raymond studied an AI assistant rolled out to more than five thousand customer support agents, productivity rose by about 14 percent on average, and by roughly a third for the newest agents. The tool worked because it had absorbed the habits of the best people and passed them on. Your experienced employees are the raw material. They should be co-authors, and they will only agree to that if they trust what happens next.

Measure the absorption, not the tool

Most AI dashboards measure the machine: tasks handled, hours saved. We put adoption next to those numbers: how many people on the team actually use the thing this month. If hours saved goes up and adoption does not, you have a demo.

The gap will get wider

The models will keep improving on their own schedule, and no company controls that. I find that exciting and, some weeks, a little unnerving. What a company does control is its own absorption rate, and that is set by trust far more than by technology. It is the one part of the gap you can close yourself.

Sources and further reading.

The research referred to in this note, for anyone who wants to read it in full.

  1. Paul David, "The Dynamo and the Computer: An Historical Perspective on the Modern Productivity Paradox", American Economic Review (1990).
  2. Erik Brynjolfsson, Daniel Rock and Chad Syverson, "The Productivity J-Curve", American Economic Journal: Macroeconomics (2021).
  3. Wesley Cohen and Daniel Levinthal, "Absorptive Capacity: A New Perspective on Learning and Innovation", Administrative Science Quarterly (1990).
  4. Writer and Workplace Intelligence, enterprise AI adoption survey (2026).
  5. Microsoft and LinkedIn, 2024 Work Trend Index Annual Report (2024).
  6. Amy Edmondson, "Psychological Safety and Learning Behavior in Work Teams", Administrative Science Quarterly (1999).
  7. Erik Brynjolfsson, Danielle Li and Lindsey Raymond, "Generative AI at Work", Quarterly Journal of Economics (2025).

More field notes

  1. Co-existence is a team sport.

    Most advice about working with AI is written for one person at a keyboard. A company is not one person.

    4 min read
  2. Everyone wants AI. Few ask what for.

    AI is brilliant at some tasks and unreliable at others that look just as easy. A plan names the task.

    4 min read
  3. When building is nearly free, the wrong thing gets expensive.

    Cheap software sends three bills that never appear in the quote. Your team pays all of them.

    4 min read

Talk to us

AI that works alongside your people.

We start with the work, and with the people who do it. The build is on us. You pay for what it delivers.

Book a call