Halfway through the quarterly review, the CEO stopped his head of digital mid-sentence. "Before the next slide," he said, "just tell me one thing. We have spent eighteen months and a serious amount of money on AI. If a competitor walked in tomorrow and asked what we got for it, what would we say?" The room reached for its laptops. Somebody mentioned the pilots. Somebody mentioned the licences. Nobody mentioned a number. The CEO let the silence sit, because the silence was the answer.
It is a scene playing out in boardrooms from Sandton to Stellenbosch, and it has almost nothing to do with whether the technology works. The models work. They are astonishingly capable and, by any historical standard, astonishingly cheap. The tools have been bought, the data scientists hired, the subscriptions renewed. Adoption is no longer the hard part. Conversion is.
The numbers make the gap impossible to look away from. BCG's Build for the Future study found that 88% of organisations have adopted AI in at least one function — but only around 6% have turned that adoption into scaled, measurable value across the business. McKinsey's State of AI puts a sharper edge on it: roughly 74% of the economic value from AI is captured by the top 20% of firms. This is not a normal spread of returns. It is a power law. The leaders are not slightly ahead. They are pulling away, because every successful deployment compounds into data, capability and confidence that fuels the next one.
The instinct, when leaders first see that gap, is to explain it with money, talent or data. We did not spend enough. We could not hire the right people. Our data is not ready. All three are comforting. None survives contact with the evidence. BCG's own research gives us the reason in a single ratio — the 10-20-70 rule. Only 10% of the value in an AI transformation comes from the algorithms, and 20% from the technology and infrastructure. The other 70% comes from redesigning how people work and how the organisation makes decisions. The firms that capture value do not, as a rule, spend more than the firms that do not. They spend differently. The 94% put 70% of their budget into technology. The 6% put 70% into people and process. Same market, same tools, opposite result.
Which means the 6% are not luckier. They are better designed.
Look at how a leader does it and the pattern becomes concrete. DBS Bank — the Singapore-headquartered group that has quietly become one of the most-cited AI organisations in the world — did not get there by buying a better model. By 2024 it was running more than 1,500 AI and machine-learning models across 370-plus use cases, and attributing over SGD 750 million in economic value to that portfolio. The headline is the number. The lesson is underneath it. DBS treated AI as a portfolio to be governed, not a collection of experiments to be admired. Every use case was tied to a value driver. The finance function co-owned the returns. The work was redesigned around the models, not the other way round. That is not luck, and it is not spend. It is architecture.
This is the move that separates the two groups, and at Green Everest it is where every engagement begins. The 6% start with three disciplines the 94% skip.
They start with value architecture. Before anyone asks what the AI can do, they ask where the value is — which revenue line, which cost structure, which risk exposure. They map the value pools and they put a rand figure and an owner against each one. The question is not "what's possible with AI?" It is the older, harder boardroom question: where is the value, and who is accountable for it?
They are ruthless about use-case selection. Activity is easy; a portfolio is hard. The 6% run every candidate through a simple filter before it earns a place: is the problem material to the P&L, can the workflow actually be redesigned, is the data available, and does the organisation have — or can it build — the capability? Four yeses, or it does not enter the portfolio, however interesting the demo. This is how 47 pilots become four value pools that matter.
And they treat it as a leadership problem, not a technology one. In the 6%, the CEO does not hand AI to the CTO and ask for an update at the next board meeting. The CEO owns the value thesis, the urgency and the resource allocation, and shows up in the rhythm that turns intent into change. AI transformation is an organisational redesign wearing a technology badge. Redesign needs an architect, and the architect sits in the corner office.
None of this requires a Silicon Valley budget or a data-science army. It requires a shift in posture — from collecting tools to engineering value, from pilots to a governed portfolio, from "what can AI do?" to "where does AI change our economics?"
If you are the leader in that quarterly review, three questions are worth putting on the table before the next one:
Can we name the three value pools our AI spend is meant to move — in rands, with an owner against each?
If we stopped every AI pilot tomorrow, which ones would the business genuinely miss, and why?
Who in this room is the architect of our AI value — not the sponsor, not the budget holder, the architect?
The honest answers tend to be uncomfortable. They are also the beginning of the work. The gap between 88 and 6 is not a gap in technology, or budget, or talent. It is a gap in design — and design is a choice a leadership team can make.
The 6% are not luckier. They are better designed.
Where does your AI value actually live — and can you defend it at your next board meeting? The AIO Compass engagement maps your value pools, aligns your leadership team, and gives you a board-ready value architecture in four to six weeks. That is where becoming the 6% begins.