Use AI to Catch What You Miss Before You Make the Call
The customer wants a delivery date. Your team has Friday on the board. The supplier’s email says Thursday. It looks like you can make the promise.
Then someone reads the email closely. Thursday is when the parts ship. Nobody has confirmed when they’ll arrive.
That’s the kind of detail I want AI to flag. I’m about to make a commitment, and one unchecked assumption could change what I tell the customer. An assistant that waits for me to ask about the ship date depends on me already seeing the problem.
What happened in the cockpit
Researchers put 22 airline pilots through a time- and fuel-constrained landing scenario in an Airbus A320 simulator. In that exercise, 20% of pilots without assistance selected the safest option. That rose to 40% when AI was available on request and to 100% in both groups where AI proactively spoke up. The group that also received an explanation showed the strongest understanding of the situation.
This was one simulator scenario with small groups. For a business owner, it raises a practical question worth testing. Where in your workflow would a well-timed alert give someone a chance to check a fact before they make a promise?
Try it on your next customer commitment
Before you send a delivery promise, price quote, or project timeline, give AI the draft and the source material your decision relies on. Use this prompt.
I’m about to make the commitment below. Flag up to three assumptions that could change what I promise. For each one, point to the exact evidence in the material I shared, explain what still needs checking, and tell me what would change if the assumption is wrong.
Then run the three questions I use in Endodyne leadership training.
- What did I believe before asking AI? I thought the parts would arrive Thursday.
- What evidence supports its recommendation? The email confirms a ship date. I still need an arrival date.
- Am I prepared to own the decision? I’ll call the supplier before I promise Friday. If they can’t confirm arrival, I’ll give the customer a time for the next update.
Start with one decision your team makes every week. At your next meeting, pull up a real example and see whether the prompt gives the person making the call something specific to verify. Ask them what they checked and whether it changed what they told the customer.