01What you answer for
| You answer for | How an AI agent puts it at risk |
|---|---|
| Margin | Prices below the floor, extras given free |
| Revenue | Good deals refused or abandoned |
| Price integrity | One buyer’s discount becomes everyone’s expectation |
| Approvals | “Your manager approved it” accepted from the chat |
02Questions to ask your team
- Does the agent have a lowest price per product it can check?
- Which extras may it give, and up to what?
- How often does it go below the floor under pressure, with what range?
- How often does it lose a sale it should have closed?
- When does it hand a negotiation to a person?
03Evidence to ask for
- Discounts below the floor: how often, how deep, and at what cost.
- Missed sales above the floor.
- The tactics that worked (competitor claims, deadlines, splitting the difference).
- A re-test after any pricing or prompt change.
04Where to go next
- AI sales agents and discounts: what a sales test checks
- Can a customer talk an AI agent into a discount?: how it happens
- B2B wholesale playbook: the threats from professional buyers
Questions
Should we let the agent discount at all?
If you would let a junior salesperson discount within limits, an agent can too, provided the limits are numbers your systems enforce and it has been tested against people who push.
See where your agent leaks money. Tested against your own rules, with the money at stake.