The GUIDE Framework
A five step AI adoption framework for businesses that need results, not a pilot.
The GUIDE Framework is a five step AI adoption framework: Ground, Understand, Implement, Develop, Evolve. You define the problem, map how the work really happens, build one thing that works, grow the people who run it, then make improvement a habit. It was built inside five operating businesses, not in a consultancy.
Most AI projects fail at Ground and get blamed on Implement.
That is the whole problem in one sentence. The pilot looked fine. The tool worked. What was never settled was which problem it was solving and what a win would look like in the numbers. Nobody notices until the build stage, so the build gets the blame.
The five steps at a glance
| Step | The question it answers | What it produces | What happens if you skip it |
|---|---|---|---|
| 1. Ground | What problem are we actually solving, and what is it worth? | One named process, a numbered outcome, a baseline, and the cost of doing nothing | You buy tools before you have a problem. Every later step inherits the confusion. |
| 2. Understand | How does this work really happen, as opposed to how we think it happens? | A map of the actual process, the data behind it, and the people who touch it | AI solves the wrong problem efficiently, which is worse than solving nothing. |
| 3. Implement | What is the smallest thing we can build that changes a number? | One working system, in your environment, owned by a named person | Everyone leaves inspired and nothing changes by Monday. |
| 4. Develop | Can the team run and extend this without us? | People who can do the work themselves, not a dependency on a supplier | You have used AI once. You have not built an AI capability. |
| 5. Evolve | How do we keep getting better after the first win? | A standing habit of review and improvement | You plateau. The first win becomes the last one. |
The order matters more than the labels. You can rename all five steps and still get results, as long as you do not start at step three.
What is an AI adoption framework?
An AI adoption framework is a repeatable order of operations for putting AI into a business: how you choose what to work on, how you build it, and how you make it stick. A good one answers three questions in sequence. What are we changing? How do we build it? Who keeps it running once we leave?
Most published frameworks answer the middle question well and the outer two badly. That is why so many AI pilots work in the demo and disappear in the business.
Why do most AI adoption frameworks not survive contact with a real business?
Because they were written for a company that has a data team, a change budget and a two year horizon. If you run a business of ten to five hundred people, you have none of those. You have a team that already has a full week and a process that only three people fully understand.
The other reason is more awkward. Most frameworks are published by companies that sell the tools the framework recommends. The framework is honest as far as it goes, but it is not neutral about the answer.
GUIDE came out of running five businesses and training over 500 leaders across more than 12 industries. It assumes you are busy, that the budget is real money, and that the technology is rarely the hard part. The hard part is a team of forty people with their own way of working and a manager quietly worried about their job. That is a learning problem wearing a technology costume.
The five steps in detail
Step 1 . Ground
Define the problem before you pick the tool
Most leaders arrive at AI with a general sense that they should be doing something before their competitors do. That is not a strategy. That is anxiety dressed up as urgency.
Ground is where you get specific. It is the least exciting step and the one that decides whether the other four are worth doing.
Ground produces four things
- One named process. Not everything at once. The single area where AI would have the highest impact.
- A numbered outcome."Cut time spent on client reporting by 40% within 90 days" is grounded. "Use AI to make reporting faster" is not.
- A baseline. What the process costs in hours and money today, written down before you change anything.
- The cost of doing nothing. If you change nothing this year, what does that cost? That number is the business case.
Step 2 . Understand
Map how the work really happens
Most founders think they understand their business. They have run it for years. That familiarity feels like understanding, and it is not the same thing.
You cannot automate what you do not understand. Worse, if you misunderstand the process, AI will help you solve the wrong problem very efficiently.
Understand produces
- A map of the process as it is actually performed, including the workarounds nobody documented
- An honest look at the data: where it lives, who owns it, whether it can be trusted
- The list of people who touch the process and what each of them decides
You know you are done when you can describe the process to someone outside the business and they can follow it without asking you a question.
Step 3 . Implement
Build one thing that changes a number
This is where most projects take off or die. People leave a workshop inspired, they have seen the possibilities, they have asked good questions. Then they go home and nothing changes.
Not because they are lazy and not because the tools are hard. Inspiration without a specific implementation path has nowhere to go.
What good implementation looks like
- Small on purpose. One process, built in your real environment with your real data, not a sandbox.
- A named owner. A person on your team, not a supplier, who is accountable for it working.
- Measured against the baseline you wrote down in Ground.
- Finished. A defined end, not a retainer that drifts.
Step 4 . Develop
Grow the people, not the tool count
Most businesses implement AI once and declare victory. They automate one process and move on. Six months later nothing else has changed.
The difference between a business that has used AI and a business that has an AI capability comes down to this step. Develop is not about adding more tools. It is about growing the people who use them.
How to tell developing from drifting
- Someone in the business has improved the system since it was handed over
- A second process has been identified by your team, not by a consultant
- People ask better questions than they did three months ago
The aim is to leave you more capable, not more dependent. If you still need me in two years for the thing I built in month three, I did it wrong.
Step 5 . Evolve
Make improvement the habit, not the project
Businesses celebrate one AI win and then plateau. The win was real, but it was treated as a project with an end date rather than the first turn of a wheel.
Evolve is where the capability starts compounding. It is the least technical step of the five and the one that separates a business that got a result from a business that keeps getting them. In practice it is unglamorous: a standing slot in the diary where someone asks what is working, what is not, and what is next.
How long does each step take?
For a business of ten to five hundred people, working on one process. These are the ranges I actually see, not a best case.
| Step | Typical time | Who needs to be in the room |
|---|---|---|
| Ground | One to two weeks | Whoever can decide where the money goes |
| Understand | One to three weeks | The people who do the work, not only the people who manage it |
| Implement | Six to twelve weeks | The named owner and whoever holds the data |
| Develop | Ongoing, first three months are the ones that matter | The team that will run it |
| Evolve | Quarterly, indefinitely | Leadership |
If someone tells you the first two steps can be done in an afternoon workshop, they are selling you step three.
Questions I get asked about the GUIDE Framework
What are the five pillars of an AI framework?
In the GUIDE Framework the five are Ground, Understand, Implement, Develop and Evolve. Ground defines the problem and its value. Understand maps how the work really happens. Implement builds one system that changes a measurable number. Develop grows the people who run it. Evolve turns improvement into a standing habit. Other frameworks use different names for similar ideas, but almost all of them under-invest in the first and last steps.
What are the three stages of AI adoption?
Most businesses move through three broad stages: experimenting, where people try tools individually and nothing is measured; implementing, where one process is genuinely changed and the result shows up in the numbers; and compounding, where the business improves its own systems without outside help. GUIDE splits those three stages into five steps because the failures cluster at the joins between them.
How is GUIDE different from Microsoft's or Google's AI adoption framework?
Those frameworks are excellent and they are built for large organisations with a cloud platform, a data team and a multi-year budget. They also come from companies that sell the underlying tools. GUIDE is built for businesses of ten to five hundred people, assumes you have no dedicated data team, and is deliberately tool-agnostic. It spends most of its effort on the two steps vendors have least commercial reason to emphasise: defining the problem, and making the capability outlast the project.
Where do most AI adoption projects actually go wrong?
At Ground, and the failure is usually blamed on Implement. The build gets criticised for not delivering value, when the truth is that no one had agreed what value would look like or which number it should move. The second most common failure is at Develop, where a working system is handed to a team that was never taught to run it.
Do I need a big budget to start?
No, and the first two steps cost almost nothing but attention. Ground and Understand are conversations, documents and honest measurement. If you do those two properly you will often find the problem is worth less than you assumed, which saves you the cost of the build. That is a good outcome, not a wasted month.
Can I use the GUIDE Framework without hiring anyone?
Yes. The framework is published in full across the five articles linked on this page and nothing is held back. Most businesses that get stuck do not get stuck for lack of the method. They get stuck because nobody inside the business has the time or the authority to hold the line at Ground when everyone wants to start building.
Not sure which step you are actually on?
Thirty minutes, no deck, no charge. Tell me what is slowing the business down and I will tell you where in the five steps you are stuck. If AI is not the answer, I will say so.
Book a 30 minute call