Question Quality Agent
Check how open and deep a rep's discovery questions were, then show them which questions to improve next time.
- Trigger: a discovery call is processed
- AI steps: rate each question and identify the two weakest
- Output: ratings on the recording and coaching suggestions sent to the rep
Agent breakdown
A blueprint to adapt, not a fixed recipe. Set the trigger, thresholds, and methodology that fit your team, and each step builds on the last, using the conversation and CRM context you already capture.
Run on discovery calls only
The agent runs when a discovery call is processed. Later calls are usually meant to be narrower, so limiting the assessment to discovery calls avoids judging the right question for the wrong kind of conversation.
- Trigger: A discovery call is processed
- Include: Discovery calls only
- Exclude: Later-stage calls
Pull the transcript and buyer replies
Next, the agent pulls the discovery call transcript, including each question the rep asked and what the buyer said afterward. The buyer's reply gives the agent the context it needs to judge whether a question opened up the conversation or produced a short answer.
- Include: The transcript from the discovery call
- Pull: Each rep question and the buyer's reply that followed it
Rate the questions
The agent rates each question as open, closed, or leading, then judges its depth by how much the buyer said in reply. It ignores logistics and scheduling questions, so the rating focuses on questions intended to uncover the buyer's situation.
It also identifies the two weakest questions and suggests how the rep could improve them next time.
You are assessing the questions a rep asked during a discovery call.
Input: the discovery call transcript, including each question asked by the rep and the buyer's reply that followed it.
For each question the rep asked:
- Quote the question exactly.
- Rate it as open, closed, or leading.
- Judge its depth by how much the buyer said in reply.
- Explain the rating using the buyer's response.
Ignore questions about logistics or scheduling. Judge the quality of each question rather than the total number of questions the rep asked.
Identify the two weakest questions and suggest a clearer or more open version the rep could use next time. Base each suggestion on the same discovery topic and do not introduce information that was not discussed on the call.
Return the rating for each question, followed by the two weakest questions and the suggested improvement for each one.
Write the rating to the call
Once the questions are rated, the agent writes the assessment to the call recording. It places the question quality rating beside the coaching score, so managers can read both together when reviewing the call.
- Update: The question quality rating
- Location: The call recording, beside the coaching score
Send the two weakest questions to the rep
Finally, the agent sends the rep a direct Slack message after the call. It quotes the two weakest questions and includes the suggested improvements, giving the rep something specific to try in their next discovery conversation.
- Recipient: The rep who ran the call
- Cadence: One direct message per call
- Include: The two weakest questions and suggested improvements
Question ratings on the recording and two quoted questions with improvements in the rep's Slack
See how open and deep the rep's discovery questions were, based on what each question prompted the buyer to say. The rating sits beside the coaching score on the recording, while the rep receives the two weakest questions and suggested improvements in a direct Slack message after the call.
AI Agents that pair well
Same data, different jobs. Run them together for a fuller picture.
Learn more about GTM & Revenue Operations
RevOps Lab Podcast
50+ Free GTM & RevOps Cheat Sheets
Free RevOps Community (1k+ members)
.webp)