How to make your planning system explain itself: five kinds of questions

By Maarten Alta · 30 October 2026

Now that planning software increasingly comes with an AI assistant, the question changes. No longer "what can the system do?", but "what do I ask it?". An assistant is only as good as the questions you ask. These five kinds of questions help you get the most out of the conversation.

1. The why question

"Why is order 4711 not planned?"

A good answer names the binding rule: a time window, the driver's driving time, the payload, or the battery. A vague answer such as "that was not optimal" is not an explanation.

Electric: "Why does truck 12 charge at 11:00 and not during the break?"

2. The what-if question

"What happens if truck 7 leaves at 6:00 instead of 7:00?"

This is where the system should recalculate. Ask that too: "Did you recalculate this?"

Electric: "What if the charger at customer X is out of order today?"

3. The comparison question

"What is the difference between this morning's proposal and this one?"

Useful when you have moved trips yourself: what does your change cost, and which rules does it touch?

4. The what-does-it-cost question

"What does it cost if we serve customer Y today after all?"

The answer must come from a calculation: extra kilometres, hours, road toll, energy. That is how you back up a choice towards a customer or management.

Electric: "What does it cost if truck 3 fast-charges on the road instead of at the depot?"

5. The pass-it-on question

"Explain to the driver in three sentences why his trip runs differently today."

Or: "Summarise for management why the cost per stop is higher this week." This is where a language model is at its strongest: the story for the right reader.

Three checks for every answer

Where does this number come from? From the optimisation model, or an estimate by the language model?
Can I find it back? The binding rule should also be on the screen.
Does it match what I know? The customer who opens later on Tuesdays is not in the data. You know that.

Practise

Asking good questions is a skill, and you learn it by doing. On the AI day of the Certified E-Fleet Planner programme you plan a real day with an AI route planner, in which all kinds of things go wrong, and you practise exactly this.