An example I made up. Illustrative figures.
1. The situation
Gopal Krishnan is in his first year as a manager. He works in the planning team of a regional chain of household goods stores, with its head office in Kuala Lumpur. The chain is choosing the site for its next store.
The regional director favours Site A in Shah Alam and will put it to the board, which meets on Thursday to decide.
2. What the AI produced
Gopal asks ShriketitAI to compare the two sites. It answers:
"Site B in Seremban earns more money by the end of year 2 than Site A in Shah Alam. Based on the two landlords' rent quotes and three months of footfall counts at each site. I am 82 percent sure."
Gopal checks where the footfall counts came from. He finds nothing wrong with them, and he believes the comparison.
3. The check
On Tuesday night Gopal drafts an email to the whole leadership team. It says the site the director favours is the weaker one. He reads it again, and he doesn't send it.
On Wednesday morning he takes the comparison to the regional director first. He shows what it rests on: the two rent quotes and the footfall counts. Then he asks one question. "What do you know about these two sites that these figures do not show?"
The director knows one thing the figures don't. Site A's landlord has offered the first year rent-free if the lease is signed by the end of the month.
4. The decision, and who was told
They agree to put both sites to the board, with both facts. Gopal presents the comparison. The director presents the landlord's offer. The board hears each fact from the person who holds it.
5. What happened after
With the rent-free year counted, the two sites are close. The board chooses Site A, with a review of footfall after six months. Three months later, Gopal's comparison is the template the planning team uses.
What the AI had, and what it didn't
The AI's comparison was sound on the figures it was given. The landlord's offer was not in those figures. Only the director had been told about it, and Gopal had been about to write to everyone else first.
The AI can compare two sites on the figures it holds. A person still has to ask who knows the facts that sit outside those figures.
Where I learned it
At twenty-nine I recorded 241 per cent growth at Electrolux. I was dismissed within the year.
What I tell them
Being right on the figures you have is where the work starts. A correct finding does nothing until the people who decide use it. So the next question for a young manager is how the finding reaches the person who will put it to them.
Who hears it first, and how, decides whether it gets used. Suppose the person who will put the proposal to the board hears it from you first. They hear it in a room, with the basis in front of them. Then they can test it and use it. If they hear it second, in an email copied to their peers, they have to defend their position before they can look at yours. The finding then gets argued over instead of used. A young manager can be right on Tuesday night and still not see the finding used on Thursday.
The person who will put the proposal to the board may know a fact the numbers don't hold. The figures in front of you are the figures someone gave the AI. A landlord's offer or a promise made to a supplier can sit with that person, and nowhere in the data. The analysis can be done well and still leave that fact out. This is why the finding goes first to the person who is responsible for the result.
So ask that person before you tell everyone. Show them what your finding rests on, and ask what they know that the figures don't. Asking first costs little time. If they know nothing more, you go to the wider group with that person beside you. If they do know more, you have found the gap before the board did. You have also given them a reason to trust your next finding.
AI can make this situation more common. A young manager can now produce a sound comparison quickly, before anyone senior has looked at the question. The AI can produce the comparison. The young manager still has to decide who sees it first, and then ask that person what the comparison leaves out.
None of this asks a young manager to stay quiet. Gopal didn't give up his finding. He presented it to the board under his own name, beside the director, and his comparison became the planning team's template. His finding was sound on the figures he had, and he made sure the people who decide could use it.
What I won't claim: that the person who will put the proposal to the board always knows more than the figures. Sometimes they don't, and the finding stands as it is.
I give keynotes, briefings and workshops on this.