Questions Your Customers Will Ask About AI

Customers have started asking about AI in security questionnaires, renewal conversations, and sales meetings. The questions are consistent, the answers exist inside the company, and the problem is that they are usually improvised by whoever is in the room. This guide covers the questions, the answers that hold up under procurement scrutiny, and the ones that create problems later.

The five questions

Question	What they need	Answer source
Does AI process our data, and which data?	Their own compliance position	Inventory and data flow mapping
Does a person review decisions about us?	Ability to contest outcomes	Autonomy and oversight design
What happens when the AI is wrong?	Remedy and error handling	Incident and correction processes
Is our data used to train models?	Confidentiality and competitive risk	Provider contracts and policy
Does AI change what we pay?	Value sharing	Commercial position, decided in advance

The first four are covered in the how to answer customer questions about AI guide, which also covers maintaining the answer set. The fifth is newer and is where most companies are unprepared.

Why has pricing become a customer question?

Because customers can do the arithmetic. A buyer paying for a service priced on labor, who knows that agents now do a share of the work, asks the obvious question at renewal: is this still priced correctly? This appears first in services, outsourcing, and managed services relationships, and it is spreading.

Three defensible positions, and the company should choose deliberately rather than improvise in a negotiation:

1. Hold price, improve service: faster turnaround, wider coverage, higher quality, with evidence. Works where the customer values the improvement.
2. Share the benefit: a price adjustment that reflects a portion of the efficiency, in exchange for term or scope commitments.
3. Restructure to outcomes: price per resolved case, processed document, or completed unit, which changes the conversation from cost to value.

The how AI agents change the unit economics of services piece covers the underlying economics; the position should be agreed by the executive team before the first renewal conversation, not after it.

Why does consistency matter more than the answer itself?

Because inconsistency is discoverable and becomes leverage. When an account manager says no AI touches customer data, a security questionnaire response describes an agent that does, and a support engineer mentions something else again, the customer's procurement team has both a trust problem and a negotiating advantage.

One maintained answer set, owned by a named person and grounded in the inventory and the actual contracts, used verbatim by account teams, with escalation for anything outside it. This is unglamorous and prevents most of the damage.

What should never be said?

Anything that cannot be evidenced:

• "No AI touches your data" when an agent processes their tickets.
• "A human reviews every decision" when review is sampled.
• "Your data is never used for improvement" when the contract permits aggregate use.
• "Our AI is 99% accurate" without an evaluation set behind it.

Each surfaces eventually in an audit, a questionnaire response, or an incident, and the customer's reaction is to the discrepancy rather than to the underlying fact. The AI reputation risk guide covers what happens when the gap becomes public.

How should contractual AI clauses be handled?

Expect them, and decide the position in advance: what the company will commit to on data use, human review, notification of AI use, audit rights, and accuracy. Commit to what is true and operationally sustainable, because these clauses bind across the term and across future deployments. A commitment that a human reviews all decisions, given once to win a deal, constrains the roadmap for years. The general counsel's guide to AI and agentic AI covers the contracting dimension.

What should executives prepare?

• The maintained answer set, with an owner and a review cadence.
• The pricing position, agreed by the executive team.
• Standard contractual language on AI that legal has approved.
• An escalation path for questions outside the set.
• A disclosure standard: what customers are told proactively.

What should executives ask?

• Could our account teams answer the five questions consistently today?
• Have we decided our position on AI and pricing before the next renewal?
• Has anyone committed in a contract to a control we do not have?
• When did the answer set last get checked against the inventory?
• What do we tell customers proactively, and is it enough?

How can FISTA Solutions help?

FISTA Solutions maintains the inventory, data flow documentation, and oversight records that customer answers depend on, through its AI enablement practice, and builds AI agents whose data handling and review points are documented and verifiable so commitments can be made safely. Since 2017, FISTA has delivered 150+ projects for 50+ companies across 12+ countries.

To prepare consistent answers before the next renewal cycle, talk to FISTA on WhatsApp, or read how to build customer trust in AI agents.