People, process, technology: why AI needs all three legs to stand
Think of an AI initiative as a three-legged chair: people, process, and technology, and it will not stand on two.
BY BETTINA MEYER··7 MIN READ
TL;DR
- People, process, technology is a framework that says a change only works when the people, the way the work is done, and the tools are all in place. For AI, I think of it as a three-legged chair: if one leg is missing, the chair will not stand, and neither will the initiative.
- In my experience, most companies have two of the three legs. The missing one is usually why an AI project "fails", or why the return you were hoping for never shows up.
- The people leg needs leadership from the leadership team, and staff who are trained and understand the value well enough to buy in.
- The process leg needs a mapped workflow and data that is ready to move through it. Of 25 attributes tested in one large survey, redesigning workflows had the biggest effect on whether generative AI showed up in earnings.
- The technology leg needs the right software, built securely and integrated with your current systems, so that it does not create new bottlenecks.
When an AI project disappoints, the first suspect is usually the technology. The model was not good enough, the vendor overpromised, the tool was the wrong one. My view is that the technology is rarely the only thing missing, and often it is the one part that was in place.
The framework I keep coming back to is an old one: people, process, technology. Below I explain what it is, why AI needs all three parts working before it pays off, what each part requires in practice, and a short checklist you can run against an initiative this week.
What is the people, process, technology framework?
People, process, technology (often shortened to PPT) is a way of checking that a change has everything it needs to work: the people who will do the work, the steps they follow, and the tools that support them. Its roots go back to Harold Leavitt's diamond model from the 1960s, and the three-part phrase became common in the 1990s through the security technologist Bruce Schneier.
I find it most useful for AI when I picture a three-legged chair. One leg is people, one is process, and one is technology. The seat is the project or use case you are trying to get across the line. If all three legs are in place, the chair stands. If one is missing, it does not matter how well made the other two are.
That last point is easy to underestimate. A chair with two excellent legs still falls over. An AI initiative with a strong tool and an enthusiastic team, but no mapped process, behaves the same way.
Why do AI projects fail when one leg is missing?
AI projects usually fail at the leg that was never built. Most companies have two of the three, and the gap stays invisible until the initiative is live and the return does not arrive. Most organisations are still struggling to generate meaningful returns from their AI initiatives, and the barriers tend to sit in people, processes, and politics rather than in the technology itself.
Each missing leg leaves a recognisable pattern.
| Missing leg | What it looks like | What to fix first |
|---|---|---|
| People | The tool is live, logins are low, and the work happens exactly the way it did before. | Visible leadership, training on real work, and a clear answer to "what does this give me?" |
| Process | The tool works in a demo but stalls in daily use, or speeds up a step that was never the problem. | Map the workflow end to end and check the data that feeds it. |
| Technology | People want to use it, but it sits outside their systems, needs copying and pasting, or raises security questions nobody can answer. | Choose software that fits the current stack and integrate it properly. |
Each of these is a design gap, and that is good news, because design is something a leadership team can change. Design gaps are also behind many of the pilots that never make it past the demo.
What does the people leg need for AI?
The people leg needs two things: leadership and adoption. Leadership means the leadership team drives the AI initiative themselves. It cannot be handed to IT or to one enthusiastic manager and left there, because people watch what their leaders spend time on and take their cue from it.
Adoption means people are trained to use the tool, and they understand the value well enough to buy in. Someone who has sat through a session but cannot say what the tool does for their own job will go back to the old way on Monday. Only about half of frontline employees regularly use AI tools, which tells me access alone does not close the gap.
This is the leg where change management for AI earns its place. Count the workflows that changed. Licences issued tell you very little.
What does the process leg need?
The process leg needs a mapped workflow and data that is ready to fit into it. Mapped means you can say exactly what happens, when, and where: who starts the work, what they hand over, which decision gets made at each step, and where it ends up.
Mapping a process feels slow next to buying a tool, so it is tempting to skip it. It is also where the value sits. Of 25 organisational attributes tested, the redesign of workflows had the biggest effect on whether organisations saw an impact on EBIT (earnings before interest and tax) from generative AI, the kind that writes text, code, and images. AI's value shows up across a whole workflow, and if one step in the chain is very hard for AI, that single step can undermine the entire operation. You only find that step if you have mapped the chain.
Data belongs on this leg too. A model can only work with what your systems can give it, so the data has to be accurate, consistent, and available at the point in the flow where the AI needs it. If a step depends on a spreadsheet someone updates when they remember, the process is not ready, however good the tool is.
What does the technology leg need?
The technology leg needs the right software, built securely, that works with the systems you already run. Four questions cover most of it.
- Is it the right tool for this task? Different models suit different work, and the cheapest suitable option is often the sensible one. It also pays to stay free to switch models as prices and capabilities change.
- Is it secure? Know what data it can see, where that data goes, and who can switch it off.
- Does it work with your current stack? A tool that lives in a separate tab becomes one more thing to remember.
- Is it integrated, or does it create a new bottleneck? If people must copy results from one system into another by hand, the tool has moved the queue somewhere else.
How to check whether your AI initiative has all three legs
- Name the leader driving it. One person on the leadership team who reviews progress and makes decisions about it. If you cannot name them, the people leg is missing.
- Ask the people who will use it what it gives them. If they cannot answer in a sentence, they need training on their own work and a clearer case for the value.
- Map the process on one page. What happens, when, and where, including handovers and decisions. Mark the steps the AI will touch.
- Trace the data into each of those steps. Check where it comes from, who maintains it, and whether it is accurate enough to act on.
- Test the fit with your systems. Confirm the tool connects to what people already use and that security has signed off.
- Look for the new bottleneck. Follow one piece of work end to end after the change and see where it now waits.
- Fix the shortest leg first. Spending more on a leg that is already strong will not make the chair stand.
Frequently asked questions
What is the people, process, technology framework?
It is a framework for checking that a change has the people, the workflow, and the tools it needs. For AI, all three must be in place before an initiative pays off. Treat them as three legs of a chair: if one is missing, the initiative falls over.
Which comes first: people, process, or technology?
My view is that technology should rarely come first. Start by naming the leader and mapping the process, because those tell you what the technology needs to do. All three have to be in place before you call the initiative live.
Why do most AI projects fail to deliver ROI?
Usually because one of the three legs is missing. Many companies have the technology and some enthusiasm, without a mapped process or adoption that sticks. The tool works, and the work around it stays the same.
Is people, process, technology still relevant for AI?
Yes, as a check on execution. It will not tell you which AI opportunities are worth pursuing, because that is a strategy question. Once you have chosen a use case, it is one of the quickest ways I know to find out why it might not land.
How do I find out which leg is weakest in my organisation?
Run the checklist above against one live initiative. For a broader view, the free AI Readiness Diagnostic scores your organisation across strategy, data and systems, people, process, and tooling.
Finding your shortest leg
If you would like a scored picture of where your organisation stands, the AI Readiness Diagnostic takes about five minutes and ends with a report you can take to your leadership team. If you would rather talk it through, our AI readiness audit looks at the same five dimensions in depth, and we are happy to start with a conversation.
WORK WITH PRAXES
Ready to move from reading about AI to getting your organisation ready for it? Start with the diagnostic.
Take the AI Readiness Diagnostic