Recognize, Try, Standardize, Automate: the four stages of team AI adoption

Where teams actually land

You rolled out the AI tools. A few people took off with them. The rest tried once, felt behind, and quietly went back to the old way of working. That is why most teams land in one of three places.

• A few have figured it out. Two or three people get real value out of it, and everyone points to them as proof it's working.

• Most don't know where to start. They feel behind, try it once or twice, and quietly go back to the old way of doing the work.

• The middle sends out AI slop. Sometimes they get a good result. Sometimes they pass along AI output nobody checked.

Leadership can't scale what it can't see. Two or three good outcomes are not a capability, and a training that faded a month later is not one either. The four stages below are the map we use to say where a team is, in one word, and what the next step is.

The four stages, one at a time

We believe all frameworks are wrong, and some are helpful. The same four stages score the readiness check, so a team that takes it lands on this map. Each stage has a plain test: what people on the team actually do, not what the tools can do.

1. Recognize. Spot where AI actually fits your real work. The test: people spot a task where AI could help before someone points it out, and they talk about where AI could improve the work.

2. Try. Test it on one real task and see what it actually does. The test: the team uses AI on real work, not just a one-off experiment, and someone tries it more than once for the same kind of task.

3. Standardize. Turn what your best people do into something the whole team can run the same way, every time. The test: the team uses a shared template, prompt, or process someone already built, and the output is consistent no matter who produced it.

4. Automate. Agents and automations, once the process is proven. The test: the team relies on an automated AI workflow to handle a repeated task, and an agent finishes work without someone driving each step.

None of the tests asks which tool you bought. They ask how often the work gets done a certain way. That is the difference between a tool problem and an adoption problem, and most teams don't have a tool problem.

Why Standardize is the biggest opportunity

The biggest opportunity sits in the gap between Try and Standardize. In the Try stage a few people get real results and the rest get uneven ones: when the team does use AI, the output is inconsistent, and some of it is slop. The know-how that produced the good results lives in two or three heads and isn't written down anywhere.

Standardize is where that changes. A shared template, a shared prompt, a process someone already built: the best person's judgment becomes something a colleague can run the same way on Monday. A shared, repeatable way of working is what turns scattered wins into output you can count on every time, and it protects the gains, because the progress stops depending on whoever happens to be in the room.

This is also the stage where leadership gets visibility. A standardized workflow can be seen, measured, and improved. A private habit cannot.

Why most teams want to jump to Automate, and why the sprint doesn't

Most teams want to jump straight to Automate. The sprint starts at Recognize and Try, and gets the team to Standardize. That's where the upside is.

The pull toward Automate is understandable. Automation is the thing that looks like progress: build the agent, replace the process, prove you are doing something with AI. The trouble is that you cannot automate what you have not standardized. An agent that runs a process nobody agreed on just produces slop faster.

AIM Recycling, one of our case studies, took the other route with a spreadsheet the business relies on. Before asking AI to take over any part of the workflow, they asked it to review the work first: check the data against known rules, flag inconsistencies, surface anything that deserves a second look. People stayed in the driver's seat, AI got a defined role, and the team could see how it performed before trusting it with more. Review before you replace. Check before you change. Read the AIM Recycling case study →

That is what proven means. Once a workflow is standardized and checked, building on it is a small step. Before that, it is a leap. After the sprint, that step is what Unlimited FDE is for. FDE stands for Forward-Deployed Engineer: a hands-on AI implementation partner who works alongside your team to turn real business processes into practical workflows inside your approved AI platforms.

Find your stage

The readiness check asks eight questions, two per stage, and takes about two minutes. Nothing in it is a grade. It shows you the current picture for your team and which stage to work on next. No sign-up, no email required.