FAQs, SOPs, Terms & Business Rules for Your AI Agent

Give your AI Agent clear answers and operating rules so it knows how your business should handle customer enquiries.

Last update on: 23 September 2026

FAQs, SOPs, Terms & Business Rules for Your AI Agent

Product information tells your AI Agent what you sell.

FAQs, SOPs and business rules tell the Agent how your business wants customer conversations to be handled.

Both are important.

Your Agent may know a product very well, but it still needs to understand what to do when customers ask about pricing, availability, delivery, quotation, warranty or special requirements.

1. Build Your FAQs

1.1 Start With Common Questions Begin with questions your team answers repeatedly.

Examples include:

  • How do I request a quotation? 
  • Do you deliver? 
  • Which areas do you cover? 
  • What payment methods do you accept? 
  • Do you provide installation? 
  • What warranty is available? 
  • What is the lead time? 
  • Can I buy or rent? 
  • Who should I contact for technical support? 

1.2 Keep Answers Clear FAQ answers should be:

  • Accurate 
  • Current 
  • Easy to understand 
  • Direct 
  • Consistent with your business rules 

Avoid unnecessary detail when a short approved answer is enough.

2. Build Your SOP

2.1 Explain What the Agent Should Do SOP tells the Agent how to handle different customer situations.

Useful SOPs may include:

  • Sales enquiry process 
  • Qualification questions 
  • Lead capture 
  • Product recommendation workflow 
  • Quotation process 
  • Technical escalation 
  • Human handover 
  • After-sales support 

2.2 Tell the Agent What Happens Next A strong SOP should not only explain what the Agent should avoid.

It should also explain the next action.

Instead of:

Do not provide final pricing.

Use:

If final pricing requires sales approval, collect the product, quantity, customer location and contact details before handing the enquiry to the sales team.

This gives your AI Agent a clear workflow.

3. Add Terms & Agreements

Use Terms & Agreements for approved commercial and customer-facing conditions.

This may include:

  • Payment terms 
  • Delivery terms 
  • Warranty 
  • Cancellation rules 
  • Return policies 
  • Rental conditions 
  • Service terms 
  • Other agreements 

The Agent can explain approved information but should not invent exceptions or make commitments requiring human approval.

4. Define Important Business Rules

Some important rules may not appear in a formal agreement.

Examples include:

  • When pricing can be shown 
  • When availability can be confirmed 
  • What customer information must be collected 
  • When sales approval is required 
  • When technical review is required 
  • Whether alternatives can be recommended 
  • What the Agent must never promise 

These rules help your Agent behave more consistently.

5. Define Human Handover

Human handover does not need to be complicated.

Simply define situations where your team should take over.

Examples include:

  • Special pricing 
  • Negotiation 
  • Final quotation approval 
  • Complex technical judgement 
  • Contractual exceptions 
  • Complaints 
  • Unconfirmed availability 
  • Questions outside the approved knowledge 

Where possible, tell the Agent what information should be collected before handover.

This allows your team to continue the conversation without asking the customer to repeat everything.

6. Avoid Conflicting Information

If two sources provide different instructions, decide which information is correct.

For example:

  • Two different warranty periods 
  • Different delivery policies 
  • Old and new pricing rules 
  • Conflicting quotation processes 

Conflicting information can reduce answer reliability.

Keep important business rules clear and current wherever possible.

7. Key Takeaway

Good AI behaviour comes from more than product knowledge.

Give your Agent clear FAQs, SOPs, terms and business rules so it knows both what to say and what to do next.

Previous Article

Product Information That Helps Your AI Agent

Next Article

AI Agent Readiness Checklist

Back to all resources

Related reading