The 70% Problem

Why AI change succeeds or fails on people and processes, and a simple test for your own plan

The AI budget fits on one slide. Licences, a line for integration, a line for the vendor’s setup fee. At the bottom, one more row: “Training: TBC.”

The meeting approves it in ten minutes. Everyone can see what the money buys. Nobody asks who will redesign the workflow, who will tell the team what changes, or what happens when the first draft from the tool is wrong.

Six months later, the tool works. The organisation does not work any differently.

The rule

In its guide for leaders, BCG describes how the companies leading on AI divide their effort. They follow what BCG calls the 10-20-70 rule: “10% of a company’s efforts should be focused on algorithms, 20% on technology and data, and the remaining 70%… on people and processes.”

Read that slowly. The algorithm, the part many AI conversations focus on, gets a tenth of the effort. The technology and data around it get a fifth. Everything else, most of the work, is about how people do their jobs and how the organisation runs.

Why the 70% gets skipped

The 10% and the 20% come with invoices, demos and delivery dates. They are easy to see, easy to buy and easy to report.

The 70% has none of that. Redesigning a workflow takes conversations. Building trust takes time. Changing what a role involves takes careful explanation. None of it arrives in a box, so it can go missing from the slide.

That is the 70% problem: the largest share of the work is the least visible, so it is the easiest to underfund.

What sits in the 70%

For an SME or a social service organisation, the 70% usually comes down to five things:

• Workflow redesign. Deciding which steps change, which disappear, and who owns the new process.
• Roles. Being clear about what each person’s job becomes, and what stays with a person.
• Skills. Enough understanding for people to use the tools well and spot when they go wrong.
• The story. A clear account of why the change is happening, what it means for each group, and what stays the same.
• Measures. Knowing, at 90 days, whether the business moved and whether people are using the new way of working.

A simple budget test

Take your current AI plan and sort every line into three columns: algorithms, technology and data, people and processes. Count time as well as money: whose hours go where?

If the third column is nearly empty, you have a tool plan. It may still be a good one. But it is not yet a transformation plan, and that gap is where the work is most likely to stall.

This test works at any size. A 15-person business and a 300-person agency will fill the columns differently. What matters is that the third column exists.

Narrative is how the 70% travels

Of the five things in the 70%, the story is the one that holds the others together. A redesigned workflow, a new role or a new skill each needs someone to explain why it matters.

That is why I lead change with narrative. I use a Story World approach: one shared story of the change, set by leadership, with room for staff, customers and partners to add their own experience. It gives people with different concerns a common language for the same change.

The 70% does not need a big programme

For smaller organisations, the 70% rarely means a large change programme. It can mean a conversation with the people who run the workflow, a one-page brief, a five-sentence story told by the right leader, and a review in 90 days.

I work by Occam’s razor here too: the simplest effort that moves people wins. What it cannot be is zero.

Where to start

The five questions in my free guide, The Question Before the Tool, sit almost entirely in the 70%. Answer them before you add another line to the budget.

Look at your own AI plan. How much of it is about the people who will use it?

rYo

---

Source

• BCG, “The Leader’s Guide to Transforming with AI”, 6 December 2024. https://www.bcg.com/featured-insights/the-leaders-guide-to-transforming-with-ai