Understand what a conversation result can tell you
Your client's latest conversation score is higher. Before you discuss progress, you'll want to know what changed in the exchange and whether the two conversations gave your client a similar opportunity to use the skill. Listening to the relevant moments helps you give that improvement a practical meaning.
Rapport Score provides repeatable measurements, timestamped evidence, and confidence checks for this review. Use them to understand how a finding was produced and how much evidence supports it. You can then explain the result in the context of the work you're doing together.
Start with a consistent measurement
The same audio produces identical measurement results in Rapport Score. The engine applies defined measurement rules before AI interprets the findings. This gives you a stable result when you return to the same recording and a consistent basis for discussing it with your client.
Repeatability answers the question of whether the measurement produces the same result again. Accuracy concerns how well that result corresponds to what happened. Validity concerns whether the evidence supports the meaning and use attached to the result. Each question deserves its own attention when you decide how much weight to give a finding.
For example, measuring a pause gives you information about timing. Deciding whether the pause helped the conversation requires examining the exchange. Understanding why your client paused may require asking. You can work through those questions without losing the practical value of the measurement.
Check how much evidence supports the finding
Rapport Score uses high, medium, low, and insufficient confidence levels based on the available conversation material. These checks consider the amount of speech from the relevant participants, the conversation segments available for analysis, and the supporting evidence items. Sessions with too little data are flagged.
If your client speaks only briefly, you have a smaller sample to examine. Review a low-confidence result alongside its supporting moments, and choose another recording if there's insufficient evidence for the skill you're working on. The confidence level helps you judge how much conversation material supports that measurement.
Recording quality and speaker assignment also matter. Clear audio, the correct focus speaker, and an appropriate conversation type give the review a sounder starting point. If the setup doesn't match the conversation, correct it before interpreting the result.
Follow AI feedback back to its sources
The platform checks AI-assisted feedback against the conversation material, including whether a quoted exchange appears in the transcript and whether a timestamp falls within the recording. It also filters unsupported claims and waits for the analysis before generating a coaching summary. Ask Coach answers questions using the platform's measurement definitions and provides citations and links to supporting moments.
You can use those links to examine a suggestion that surprises you. Listen to the exchange and consider the surrounding conversation. Your client may have useful context about a previous agreement, a time constraint, or the purpose of the question. Bring that information into the coaching decision.
Explore Evidence, Interpretation & Inference
Compare progress in a way that helps your client
Choose conversations that give your client a reasonable opportunity to use the skill you're developing. If you're working on discovery questions, comparable discovery calls will usually be more informative than comparing a discovery call with a presentation. Look at the relevant behavior and recorded moments alongside the score.
For training follow-up, agree on what you'll examine before the review. You might look for a taught skill across several conversations, discuss where it was used, and identify the next practice task. This gives your client a clearer account of how learning is being applied. Connecting that development to revenue, retention, or another business result requires information about those outcomes as well.
Rapport Score's measurements describe communication within the available recordings. They don't establish a person's motives, honesty, diagnosis, or complete emotional state. Use the feedback to choose a useful next attempt and review progress across conversations.
Start with a recording you know well. Explore the result, listen to the evidence, and select one skill to work on. That first review will help you see how the platform can support your own approach to development.