Cost and ROI

The 10-20-70 rule: why process change beats technology

Ten percent algorithms, twenty percent technology and data, seventy percent process change. That ratio is the difference between AI that works and AI that does not. Most boards have it backwards.

By Zakir Hoosen, Director, Fuzzelogic Solutions. Board-level guidance in plain English.

Most boards think AI value lives in the technology. Buy the right tool, hire the right vendor, press the button, save the money. That is not how it works. The technology is the smallest part of the equation, and the part most teams over-invest in.

The 10-20-70 rule comes from BCG's research on what separates companies that get real value from AI and those that spend a lot and get very little. It is one of those findings that sounds obvious once you hear it, but almost nobody acts on.

This article explains the rule, what it means for your budget, and why the seventy percent is where most boards are looking in the wrong direction.

What the rule says

BCG found that the value of AI breaks down into three parts.

1. 10% comes from algorithms and models.
2. 20% comes from technology and data infrastructure.
3. 70% comes from process change, people, and ways of working.

Seventy percent. That is where most of the value lives. It is also where most of the cost is underestimated, where most of the risk sits, and where most boards spend the least time.

The algorithm is the clever bit. The technology is the visible bit. The process change is the bit that actually makes it work. And it is the bit that most projects budget for last, if they budget for it at all.

Why this matters for your budget

When a vendor presents an AI project, the price they quote is usually the twenty percent. The software, the infrastructure, the licences, the hosting. That is the part they can price because it is the part they sell.

The ten percent is often bundled in. A data science team builds a model, tunes it, and delivers it inside the twenty percent. It is clever work. It is also the part that is most likely to work, because it happens in a controlled environment with clean data and smart people.

The seventy percent is the part nobody can price because it is the part that happens inside your business. It is the workflow redesign. The team retraining. The change in how decisions get made. The politics of who loses authority when a machine starts making calls they used to make. The conversation with the union. The rewrite of the job descriptions. The six months of teething problems while people learn a new way of working.

That overrun is almost always in the seventy percent. The technology costs what the vendor said it would. The process change costs three times what anyone estimated, because process change is hard, unpredictable, and deeply human.

What IBM got right

IBM is one of the few companies that has publicly stated the scale of value from getting this ratio right.

Four and a half billion dollars. That is not a technology story. IBM already had the technology. The value came from redesigning how work happened, at scale, across a massive organisation. The technology enabled it. The process change delivered it.

That is the lesson. The technology is necessary but not sufficient. Without the process change, you have an expensive tool sitting on a shelf. With it, you have a business that works differently and better.

Where boards go wrong

The pattern I see most often is the board that approves the technology budget and then is surprised by the change budget. The project starts, the technology is delivered on time and on budget, and then the real cost begins.

The team discovers that the data is not where it needs to be. The workflow does not fit the new system. The people who have to use it were not consulted and do not trust it. The manager whose team is most affected was not in the room when the project was approved and is now quietly sabotaging it.

None of these are technology problems. They are all process problems. And they are all predictable, which means they should be budgeted for, planned for, and governed.

Rushing is the enemy of the seventy percent. You can buy technology quickly. You cannot change how a business works quickly. The teams that try end up with expensive technology imposed on unchanged processes, which is the most expensive outcome of all.

How to apply the rule

The 10-20-70 rule is not just a finding. It is a planning tool. Use it before you approve any AI project.

1. Ask what the process change is. Not the technology. What workflow changes, who does it differently, and what training is required.
2. Cost the process change separately. If the project cannot tell you what the seventy percent looks like, it is not ready.
3. Plan for the process change first. The technology follows the process, not the other way around.
4. Measure process adoption, not technology deployment. A system that is live but unused has zero value.

The fourth point is the one that matters most. A system that is deployed but not adopted is an expensive mistake. Adoption is a process problem. It is about whether people trust the system, understand it, and find it useful in their daily work. If they do not, the technology is irrelevant.

The honest version

The 10-20-70 rule is uncomfortable for technology vendors because it says that seventy percent of the value comes from work they do not do and cannot sell. It is uncomfortable for boards because it says that the hardest part of AI is not the technology. It is the people and the processes, which are the bits nobody can put on a slide deck.

That is also why it is useful. If you know where the value lives, you know where to focus. If you know where the risk lives, you know where to govern. And if you know where the cost lives, you know where to budget.

Fuzzelogic has spent nineteen years modernising banking, insurance, healthcare, retail, manufacturing, and government platforms. We tell boards what most consultants will not: if the honest answer is that AI should not touch a process, we put it in writing rather than build it anyway. The strategy worth approving is the one that understands the seventy percent and plans for it honestly.

Start with the assessment. Two to four weeks, fixed price, and you own the verdict and the roadmap whether or not we build any of it. When you are ready to talk AI, call Fuzzelogic Solutions and ask for Zak. www.FuzzelogicSolutions.com | info@FuzzelogicSolutions.com | +44 (0)1624 618950

For a practical walkthrough on measuring value before you spend, read AI ROI: how to measure value before you spend. For what AI actually costs, read How much does AI implementation cost. The full library is on our index. Our site explains how Fuzzelogic approaches AI for business. You can reach Zak directly via our contact page.