After every completed call, Sam extracts structured data from the conversation: scores, follow-up actions, quotes, and a retention risk signal.

Sam's results live in two places:

Every Sam call is logged in the Responses tab. At the top, narrow the list by survey (a dropdown that defaults to All) and by date range (defaults to All time).

Four summary cards sit above the table and double as status filters, each showing a live count:

The table lists one row per call, with columns for Evaluation (Positive, At risk, or Follow-up required), Follow-up Urgency (Low, Medium, High, or Immediate), Talent (the worker), Satisfaction Score (a 1–5 rating shown as filled dots), Survey (which check-in it was), Summary, Outcome (Completed, No answer, or Voicemail left), and Date. Use the Columns control to choose which columns appear, and click the Date header to sort.

Click any row to open the call detail. It opens in a panel headed with the check-in title (for example, "First week check-in summary"), the date, and the call duration, followed by a recording player: press play to listen, adjust the volume, or change the playback Speed.

Below the player, a Conversation outputs section lists everything Sam extracted from the call as labeled fields, such as:

At the bottom, the full Transcript shows the conversation as threaded, timestamped Agent (Sam) and worker turns, with an Original toggle. A panel on the right shows the worker's information (name, job, location, contact details, and Worker ID), so you have context while you review.

The Insights tab aggregates completed feedback so you can see recurring themes instead of reading call by call.

Three controls at the top scope what you see:

At the top of the report, a Summary gives an AI-generated rundown of the most urgent issues and what's working, with a "Generated by AI" timestamp. Use Regenerate to rebuild it. If you pick a period whose report hasn't been built yet, Sam generates it on demand, and you'll see "Your report is being generated. This may take a few minutes…" while it works.

Below the summary, insights are grouped by evaluation (Negative, At risk, Follow-up required, and Positive), each with a count. Every insight card shows topic tags (for example, Scheduling, Job readiness, Pay & benefits, or Communication quality), a title, a short description, a representative worker quote, and how many mentions it drew across calls. Each card has two actions: Follow up (open it in Cue) and Details.

Click Details on any insight to see the individual calls behind it. The table lists one row per call, with columns for Evaluation, Follow-up Urgency, Talent, Satisfaction Score, Survey, Summary, Outcome, and Date. Use the Columns control to choose which columns appear.

Sam produces several blocks of structured output from every completed conversation.

An overall satisfaction score on a 1–5 scale captures how the worker feels overall:

Every score comes with reasoning that explains why it was assigned, anchored to specific things the worker said.

Each check-in also produces a milestone-specific experience score on the same 1–5 scale, for example a Day 1 experience score or a 30-day experience score. A blank (null) score means the topic was never raised on the call; it is not a neutral score.

When a call surfaces something that needs attention, Sam captures whether a follow-up is required, its urgency, the owner who should handle it, and a self-contained note describing what needs to happen.

Urgency runs Low (can batch) → Medium (within days) → High (before the next shift) → Immediate (safety or intent to leave, act the same day).

Every execution records an outcome: Completed (a meaningful conversation occurred), Voicemail left, No answer, Callback scheduled, or Worker has left the role. When the outcome is a voicemail or no-answer, all score fields are blank.

Sam sets a retention intervention flag when it detects at-risk signals: doubt about continuing, a satisfaction or experience score of 2 or below, an unresolved concern about pay, safety, or scheduling, or a significantly negative tone. Sam is calibrated to over-flag rather than miss a worker at risk.

Most check-ins extract a notable direct quote from the conversation. Quotes are useful for sharing real worker voices, in a report or with a manager, without sharing the full transcript.

Each insight has a Follow up button that opens Cue in a side panel and asks it to "Analyze this insight and give me actionable items," with the insight's context already attached. Cue returns a What's happening read on the pattern and a numbered list of Actionable items, then offers a next step, for example "Want me to pull the call records for these workers so you can see who needs outreach first?"

You can also open Cue from Ask Cue (top right) and ask it directly about anything Sam has gathered. Cue opens with a few starter prompts (for example, "Analyse past 7 days completed Sam calls") and pulls from all completed calls to answer. Ask about satisfaction scores across touchpoints, the main pain points, or trends over time, and Cue can respond with charts and tables, for example average score, completed calls, follow-ups flagged, and retention interventions by touchpoint, or a pain-points table with a risk level per touchpoint.

Results can be exported (CSV) or opened Fullscreen, and Cue offers suggested next steps such as Break down scores by location, Show me the workers at highest risk, or Create a Report.

Cue can also escalate on your behalf. If a worker has a supervisor or a location owner, ask Cue to escalate and it will find that person and send them a message with the context of the conversation.