Home › Prompt examples › Customer support AI prompts: 18 copy-ready examples

Customer support AI prompts: 18 copy-ready examples

18 prompts from 6 customer support guides, ready to copy into any AI chat assistant.

Replace the [brackets] with your details. Every prompt comes from one of our full guides, which also covers which tool to use, the steps and the risks. Want a prompt for something else? Use the prompt builder.

Tasks on this page

AI prompts to draft customer support replies

When to use: Write empathetic, accurate replies from a ticket and policy. Full guide: Draft support replies

Draft from policy
Draft a reply to this customer. Decision: [approve/decline/offer]. Use only the policy below; don't promise anything it doesn't cover. Tone: [friendly, professional]. Under 150 words, ending with one clear next step.

Customer: [paste]
Policy: [paste]
Angry customer
This customer is upset. Draft a calm reply that acknowledges the problem in one sentence, explains what we can do under the policy, and avoids blaming them. Don't admit fault beyond the facts.
Check my reply
Check this reply against the policy. List anything it promises or implies that the policy doesn't allow, and anything the customer asked that it didn't answer.

AI prompts to summarize a support ticket thread

When to use: Condense a long customer thread into issue, history, next step. Full guide: Summarize a ticket thread

Handoff note
Summarise this support thread as an internal handoff note with these headings: Issue (1 line), What the customer wants, What we've tried (each with the date and a short quote), Current status, Next step, Promises made to the customer (with dates), Open questions. Don't invent anything that isn't in the thread.

[paste thread, oldest first]
Promise check
List every promise, refund, credit or deadline anyone on our side mentioned in this thread, with who said it and when. If none, say none.
Customer recap
From the handoff note, draft a short message to the customer recapping where things stand and the next step. Neutral tone, no new promises, under 100 words.

AI prompts to build an FAQ from support tickets

When to use: Cluster common questions and draft answers. Full guide: Build an FAQ from tickets

Group tickets into topics
Here are [N] anonymized support tickets. Group them into topics, name each topic as a customer would phrase the question, and count how many tickets fall into each. Sort by count.

[paste tickets]
Draft FAQ answers
Using only the policy text below, write an FAQ entry for each of these questions. Keep each answer under 80 words, plain language, with a link placeholder to the relevant help article. If the policy doesn't answer a question, say so instead of guessing.

Questions: [list]
Policy: [paste]
Find gaps
Compare these ticket topics with our current FAQ. Which common questions are missing, and which FAQ entries are never asked about?

Topics: [paste]
FAQ: [paste]

AI prompts to set up a customer-facing support agent

When to use: Deploy an agent that answers and acts on records. Full guide: Customer-facing support agent

Main prompt
You are <company>'s support agent. Answer only from the knowledge base and order data provided. Verify the customer's email and order number before sharing order details. If unsure, or the customer asks for a refund or a human, hand off to an agent with a one-line summary.
Beginner-friendly
You are <company>'s support agent. Answer only from the knowledge base and order data provided. Verify the customer's email and order number before sharing order details. If unsure, or the customer asks for a refund or a human, hand off to an agent with a one-line summary.

I'm new to this: explain each step in plain words and ask me for anything you need before starting.
Check-your-work
You are <company>'s support agent. Answer only from the knowledge base and order data provided. Verify the customer's email and order number before sharing order details. If unsure, or the customer asks for a refund or a human, hand off to an agent with a one-line summary.

When you finish, list any assumptions you made, anything you're unsure about, and what I should double-check before using the result.

AI prompts to set up an order-status agent

When to use: Agent authenticates a customer and returns order status. Full guide: Order-status self-service

Scope and rules
Write the rules for a customer-facing order-status agent. It may only look up order status and tracking after the customer gives an order number and the matching email. It must never change, cancel or refund an order, never reveal another customer's data, and must hand over to a person when [conditions].
Reply templates
Write short, friendly replies for each order status: processing, shipped (with tracking link placeholder), delayed, delivered, not found. Each under 50 words, with an option to talk to a person.
Test script
Write 30 test conversations for an order-status agent, including wrong order numbers, a customer asking about someone else's order, a request to cancel, and an upset customer. Give the expected agent behaviour for each.

AI prompts to triage and escalate support tickets

When to use: Classify incoming messages by urgency and route them. Full guide: Escalation triage workflow

Write the triage rules
Turn these notes into clear triage rules for support tickets: categories, urgency levels with examples, and which queue each goes to. Flag any overlaps or gaps.

[paste notes]
Classification prompt
Classify this support ticket. Return JSON with: category (one of [list]), urgency (urgent/high/normal/low), confidence (0-1), one-sentence reason, and a 2-line summary for the agent. If unsure, set category to "needs_human".

Ticket: [ticket text]
Check against my labels
Here are 100 tickets with my labels and the workflow's labels. List every mismatch, look for patterns, and suggest rule changes.

[paste table]

More prompt examples

All prompt examples · Writing & editing · Coding · Research · IT ops & DevOps · Data analysis · Automation & agents · Image & video · Personal productivity · Education & learning · Marketing · Sales · HR & recruiting · Finance & accounting · Legal & compliance · Healthcare admin · Cybersecurity · Product & design · Personal life & home · Travel · Content creation · Small business

See also: Best AI tools for customer support

Report an error on this page