Reviews & Coaching
The AI grader gives every attempt a rubric score in seconds. A human coach turns that score into improvement. Reviews are the place coaches add written feedback, drop comments on specific lines of the transcript, and decide whether the attempt counts as passed.
What a review contains
| Field | What it does |
|---|---|
| Pass / fail | The coach’s verdict, regardless of the AI score. |
| Per-criterion adjustment | Bump the AI’s score up or down with a one-line reason. |
| Overall feedback | A short paragraph addressed to the trainee. |
| In-line comments | Comments anchored to specific transcript turns. |
| Coaching action | Optional follow-up — schedule a 1:1, assign a remedial scenario, mark for re-attempt. |
How review work fits into the week
A typical supervisor:
- Daily — skim the AI score dashboard, open any attempt that scored below threshold or above a “too good” outlier.
- Weekly — review one or two attempts per direct report regardless of score (sampling).
- Per cohort — calibrate with the AI grader to keep scores honest.
You don’t need to review every call. The AI grader handles the volume; coaches review for signal — the outliers and the sample.
Add a review
Open the attempt
From Training → Attempts or from the trainee’s profile, open the attempt. The transcript, audio, and AI score appear side-by-side.
Listen / read
Play the audio and read the transcript. Pause to drop in-line comments on specific turns:
- ✅ “Great acknowledgment here.”
- ⚠️ “Could have asked the discovery question sooner.”
Adjust the rubric (if needed)
If the AI scored something off, click the criterion, set your number, and add a one-line reason. The trainee sees both the AI’s score and yours.
Write overall feedback
Two or three sentences. Be specific: “You handled the policy explanation well — work on the open question at the start.”
Set the verdict and next step
Pass / fail / conditional pass, and an optional follow-up action.
Coaching actions
| Action | What it does |
|---|---|
| Schedule 1:1 | Books time on your calendar via the configured calendar integration. |
| Assign remedial scenario | Adds a specific scenario to the trainee’s path with priority. |
| Mark for re-attempt | The trainee re-runs the same scenario. The original attempt stays in the history. |
| Promote to live | Marks the trainee certified for the skill the path covers. |
Coaching at scale
For larger teams:
- Coach assignments — auto-route each attempt to a specific coach based on team or skill.
- Coaching SLA — flag attempts older than X days that haven’t been reviewed.
- Coaching dashboard — per-coach review velocity and consistency.
Don’t fire-and-forget the AI score. Trainees can tell when nobody’s reading their work — the coaching layer is what makes training stick.
Open in Omniflow
Related
| If you want to… | Go to |
|---|---|
| Customize the rubric | Scoring Rubric |
| Read the AI’s per-call insights | AI Insights |
| Build a remedial scenario | Scenarios |