AI agent for call center supervisors
Call Handle Time Outlier Review Agent
A weekly list of handle time outliers with the likely causes grouped and the right action proposed for each
What it does
A supervisor sees average handle time creeping up 40 seconds and cannot tell whether agents need coaching, a tool is slow or a new policy confuses callers. Each week this agent reads handle times by call type and by agent, finds outliers against the normal range for that type, and reads the call notes and tags for each one to look for causes, such as a system lag, a missing knowledge article, a billing change or a repeated customer question. It groups causes and checks whether they repeat across agents and days, since a repeated cause points to a system or process issue and a cause tied to one agent points to coaching. It then drafts the right follow-up for each group. The supervisor approves any coaching or system ticket. Edge case: long calls from a known outage are labeled and excluded from coaching.
How it works
Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.
Read the steps as a list
- Week closes
- Read call records and compute the median handle time per call type
- Find outlier calls and exclude those tied to known incidents
- Read the notes and tags for each outlier
- Assign a likely cause and group similar causes
- Does each grouped cause repeat across enough calls to be meaningful?If not: Check further call notes or move the call to a miscellaneous list. Back to step 4.
- Classify each group as a system, process, knowledge or coaching issue
- Draft the follow-up for each group, such as a ticket or a coaching topic
- Supervisor approves each coaching action or ticketThe agent waits here for your OK.
- Did the following week's outliers in these groups decline?If not: Reopen the group and suggest a different action. Back to step 3.
- Outlier report
How it decides
An outlier is a call more than twice the median for its type. A cause that repeats across 3 or more agents is treated as a system or process problem, otherwise as coaching.
- An outlier is a call over twice the median for its type
- A cause repeating across 3 or more agents is a system or process issue
- Exclude calls linked to known incidents from coaching
- Never rank agents on a single week's outliers
Make it yours
Every agent is a starting point. You choose these settings for your own situation.
- Outlier multiple (default 2 times median)
- Call types
- Reporting day
- Cause categories
- Number of agents that marks a system issue (default 3)
What keeps you in control
It always asks you first
- Supervisor approves coaching actions and system tickets
Hard limits
- Never decide on discipline or performance ratings
- Keep call content within the center's own systems
It stops when
- Done: Each outlier group has a cause and an approved action
- Stop: Notes are too thin to find causes, so the agent asks for tagging improvements
Set it up
We guide you through the set-up, step by step
Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.
- One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
- The agent then walks you through connecting your own data, one source at a time
- A downloadable copy with the flow chart, the rules and the full guide