Asking well · for customer support professionals
Asking well in customer support
A request such as "reply to this customer" leaves the model to decide what the reply should achieve. It may apologize at length, restate the problem, and close without a next step. The customer then writes back, and the ticket stays open.
Asking well means stating what an experienced agent would decide before typing. What is the customer asking for? What can we offer? What should happen after they read the reply? What must the reply not promise?
Where it comes up
A reply to a billing dispute
A customer says they were charged twice. A strong request states what the account shows, what you can offer, and the next step for the customer. It also names what the reply must not promise, such as a refund date you do not control.
A macro for a recurring question
The team answers the same password-reset question many times a week. A good request asks for a macro with placeholders for the customer’s name and product plan. It sets a length limit and asks for one version for email and one for chat.
A reply in the customer’s language
A customer writes in Portuguese and nobody on shift reads it fluently. Ask for a translation of their message first, then a reply in Portuguese with an English version beside it. You then review the English before sending.
What it looks like at each band
E1 · Foundational
Gets usable drafts on routine tickets when the request is short and the issue is simple. Requests name the topic but leave out the outcome, the offer, and the channel. Fixes the draft by hand rather than by improving the request.
E2 · Proficient
Writes requests that state the customer’s question, the resolution on offer, the channel, and the length. Asks for the next step to be explicit. When a draft misses, adds the missing detail to the request instead of starting again with a vague one.
E3 · Distinguished
Splits complex tickets into steps: identify each question the customer asked, then draft an answer to each, then assemble the reply. States what the reply must not contain, such as blame or speculation about a bug. Asks the model which questions in the ticket remain unanswered.
E4 · Exceptional
Writes the request templates the team uses for common ticket types, with fields for account facts, the offer, and policy limits. Knows when a ticket needs one precise request and when it needs a short exchange. Other agents improve by copying how this person asks.
Common mistakes
Asking for "empathetic" without saying what that means
The model reads this as a request for more apology. Say what you want instead: acknowledge the specific problem once, then move to the fix.
Leaving out what you can offer
If the request does not state the resolution, the model invents one. It may offer a credit, a callback, or a timeline that nobody approved.
One request for a ticket with three questions
Customers often ask several things in one message. A request that does not list them gets a reply that answers the first and ignores the rest.
A worked example
A reply to a customer whose annual plan renewed after they meant to cancel.
Write a reply to this customer about their renewal. Be empathetic.
The model does not know whether a refund is possible, what the policy says, or what the customer should do next. It will write a long apology and may promise a refund the team cannot give.
Write an email reply to a customer whose annual plan renewed four days ago. They meant to cancel. Our policy allows a full refund within 14 days of renewal, so they qualify. Confirm the refund, say it reaches their card in five to ten business days, and tell them their access ends today. Acknowledge the frustration in one sentence. Under 120 words. Do not offer a discount to stay.
The request states the facts of the account, the policy, the resolution, the timeline, and one limit. The reply can be short and correct because the request is specific.
The full study guide for this skill, for every role, is Asking well. The other five skills for customer support professionals are on the AI skills for customer support professionals page. A one-page summary is in the AI prompt checklist.
See where you stand
The free check grades asking well and the other five skills in about fifteen minutes. Nobody sees the grade but you.
Then build the skills and prove it. The three steps for customer support professionals list the courses and templates that fit the work.