Flag conversations for coaching

Flagging is Monitor’s coaching primitive. You spot something worth discussing — good or bad — flag it with a comment, then revisit it later in a 1:1 with the agent. The flag persists; it doesn’t get lost in the daily noise.

Done well, flagging is the difference between coaching with abstract data (“your CSAT is down 5%”) and coaching with specific examples (“this conversation last Tuesday — let’s look at what happened around message 3”).

Before you start

• Supervisor permissions on Monitor
• A Monitor view scoped to the conversations you’re supervising (see Create a Monitor view)
• A coaching cadence — most teams do weekly 1:1s with each direct report

Steps — flag a conversation

1. Open Monitor, with your supervision view selected.
2. As you watch the live view, when a conversation catches your attention (positively or negatively):
   Click the conversation card to open the right-side detail panel
   Read enough to confirm what you noticed
   Click the Flag action (typically a flag icon in the conversation header or card menu)
3. Write a comment explaining what you noticed:
   Specific is better than vague: Great de-escalation in messages 4-6 — used customer's name, acknowledged frustration, offered concrete options beats nice job
   Time-stamp if possible: tone shift in message 3 — felt dismissive
   Note the action you’d want: consider escalating earlier next time
4. Save the flag.

The conversation card now shows a colored flag icon. The flag persists across status changes — a flagged Done conversation stays flagged.

Flagging principles that pay off

Flag good work as often as bad

The temptation is to only flag problems. Don’t — agents notice when their flag log is all critique. A 70% positive flag rate creates a coaching environment where feedback feels like growth rather than punishment.

Be specific about what

A flag with a vague comment is barely better than no flag — by the time the 1:1 arrives, neither of you remembers what specifically was noteworthy. Force yourself to point at a moment.

Flag in the moment

Flagging at the moment you notice catches details you’ll forget by end-of-day. A 30-second pause to flag is more valuable than a Friday-afternoon attempt to remember the week’s standout conversations.

Don’t over-flag

If you flag every conversation, the flag stops meaning anything. Aim for ~5–15 flags per agent per week. More than that and the 1:1 conversation becomes overwhelming.

Steps — review flags during a 1:1

For each direct report, weekly or bi-weekly:

1. Open Monitor → filter to flagged conversations for this agent
2. Walk through 5–10 of the flags together
3. For each: open the conversation detail, read the moment together, discuss what worked or could improve
4. Decide on one or two action items the agent will focus on next week
5. Mark the flag as Reviewed (or unflag) so it drops out of the queue

The action items become next week’s flag focus — “I’m going to flag any conversation where you handle a refund question, so we can compare your approach over time.”

Steps — agent-visible flags (when configured)

Tenants can configure flags to be visible to the flagged agent. When this is on:

• The agent sees a notification when a flag is added
• They can read the comment immediately
• They can leave a response on the flag (acknowledgment, question, context)

This creates a continuous feedback loop without waiting for the 1:1. Good for fast-iterating teams; some teams prefer flags to stay private until the 1:1 conversation.

The visibility is a tenant-wide setting in Monitor configuration.

Patterns that work

• One-week review — flag through the week, review all together on Friday — Stable teams with predictable rhythms
• Just-in-time — flag and discuss within 24 hours when it’s a clear teaching moment — New agents in their first 30 days
• Theme weeks — focus flags on a specific topic (de-escalation, escalation timing, tone) — When working on a specific skill
• Positive recognition — flag good work and share it across the team — Building a culture of noticed-good-behavior

Anti-patterns

• Flagging without comments — just a flag icon with no context is barely useful in retrospect.
• Flagging only after CSAT scores arrive — the 1:1 then becomes “explain this score” instead of “look at the actual conversation.”
• Public-shaming flags — when visibility is on, flags should be coaching, not punishment. If something needs disciplinary discussion, that’s a separate conversation, not a flag.