An example I made up
The company, the people, the numbers and the dates below are all made up.
The situation
Benedict Woo is head of sales at a software company in Singapore. It sells booking and scheduling software to clinics in Singapore, Malaysia and Thailand. New business has been coming in faster than anyone planned for. The chief executive wants the hiring plan for the next half year, and the first new people would have to start before 1 March.
What the AI produced
The company's planning assistant, AntbirdAI, drafts the plan. Its summary line reads:
"Hire 10 more sales staff before 1 March. New business in the pipeline grew 3 times in two quarters. I am 86 percent sure."
The pipeline figure is correct. The pipeline really did grow 3 times in two quarters. The plan follows directly from it: more demand, so more people to sell.
The check
Before he sends the plan to the chief executive, Benedict asks one question of someone outside sales. He goes to the head of onboarding and asks how long a new clinic now waits before it can start using the software.
The answer is 7 weeks. A year ago it was 3. And 2 clinics that signed last quarter have asked to cancel before they even started.
The plan was built on the pipeline. The onboarding queue was not in it, so someone still had to ask what happens to a clinic after it signs.
The decision, and who was told
Benedict decides to hire 4 sales staff and 3 onboarding staff now. On 1 June he will decide whether to hire the other 6 sales staff. He will consider them only if the wait is back to 4 weeks or less by then.
He tells the chief executive first. Then he tells the finance director, the sales team and the onboarding team. He explains to each of them why his plan is smaller than the one the AI proposed, and what would let him consider the other 6 hires.
What happened after
Three months later the wait is down to 4 weeks, and no clinic has cancelled before starting. In June, Benedict approves 4 of the 6 remaining hires.
Because the part he held back had a date and a condition, the decision in June had evidence behind it. Nobody had to argue the whole plan again from the start.
Where I learned it
I was Chief Marketing Officer at entomo from 2018 to 2020. In that period the company's yearly recurring revenue grew from 4 million to more than 25 million US dollars within 12 months.
The example above isn't from entomo. I set it in a different business on purpose. What I learned is about how decisions behave when a business grows that fast.
What it taught me
Growth makes decisions arrive faster than the evidence for them. A plan is based on last quarter's numbers, and the next decision is due before this quarter's numbers are ready. If you wait for complete evidence, the date still arrives and the decision gets made without it.
So the useful step is to split the decision. Decide now on the part that is urgent, or that is easy to reverse if you turn out to be wrong. Then name the part that can wait, and give it a date or a condition. Benedict decided to hire 7 people now and gave the other 6 a date and a condition. Postponing a named part until a named date is still a decision, and everyone involved knows when it will be made and on what evidence.
Postponing without a date is different. It's only waiting, and nobody knows what would end it.
Growth also taught me that a number that is growing fast still has to be checked by someone who knows what the business can deliver. Benedict's pipeline figure was right. What it didn't show was whether the company could serve the clinics it was signing. The person who knew that worked in onboarding, not in sales. The plan changed only because Benedict walked over and asked.
That is a habit worth building into any growth plan. Before the plan is sent for approval, find the person who sees what happens after the sale, and ask them one plain question. How long does a new customer wait now, and is that getting better or worse?
Where AI fits
With an AI planning assistant, this matters more. The plan arrives sooner and comes with a confidence figure. "I am 86 percent sure" sounds settled. Sounding right is a style, not a signal.
AI can speed up existing work. People still decide what needs to change. In the example, the AI drafted the hiring plan. Benedict still had to decide that onboarding needed to change before sales could grow. He still had to tell the people the decision affects.
When the number is growing fast, the questions to ask are simple ones. Which part of this decision is urgent? Which part can I reverse? Which part can wait, and until when? Who sees the part of the business this plan leaves out?
What I won't claim
The example is made up. The growth at entomo was the company's, and I don't claim it as mine alone. I don't measure anyone's decisions, and nothing here grades how well a leader decides.
I give keynotes, briefings and workshops on this.