Proven Manager Enablement & Coaching We built and ran this one

Manager Coaching Agent

A manager coaching agent is an AI agent for manager enablement & coaching that sits in Slack or Teams and coaches a manager through whatever they are facing right now: a report who has gone quiet, a peer conflict, a promotion request they cannot grant, a team that missed a deadline. Asks the questions a good coach would ask, offers frameworks and options, and helps the manager settle on an approach and the words to use.

How does the manager coaching agent work?

What flows in, what the agent does with it, where a person decides, and what comes out.

  1. Reads from

    Manager's situation, in chat · Coaching and skills framework · Company values and policies · No HR records or performance data

  2. AI agent · runs when a manager brings a live situation to the chat

    Manager Coaching Agent

  3. A person sets the rules and can override

    Manager decides what to do; L&D sets scope and privacy rules

  4. Produces

    Private coaching conversation · Concrete plan and wording · Optional follow-up check-in

What does the manager coaching agent do?

Sits in Slack or Teams and coaches a manager through whatever they are facing right now: a report who has gone quiet, a peer conflict, a promotion request they cannot grant, a team that missed a deadline. Asks the questions a good coach would ask, offers frameworks and options, and helps the manager settle on an approach and the words to use.

What does it produce?

A private coaching conversation ending in a concrete plan, plus optional follow-ups that check how it went

Who decides?

The manager decides what to say and do; the agent asks, suggests approaches, and drafts. Coaching conversations are private to the manager and are never shared with their own manager or HR. Only aggregate capability trends leave the system. Conversations are not retained as HR records.

What systems does the manager coaching agent connect to?

Examples of the kind of systems this agent would read from or write to, so you can picture it in your own stack. The actual set is whatever you run.

  • Chat

    where the coaching happens

    SlackMicrosoft Teams
  • Knowledge / policy

    values and policies for grounding

    ConfluenceSharePointNotionGoogle Workspace
  • Identity

    who is a manager, nothing more

    OktaMicrosoft Entra ID
  • Reporting

    aggregate capability trends only

    Power BILooker

What data does it need?

  • a coaching and skills framework
  • company values and policies for grounding
  • chat platform integration
  • nothing from HRIS or performance records

How would you measure it?

weekly active managers and return rate, by function; manager-rated usefulness of plans, per session; aggregate topic mix, quarterly, above minimum group size; manager confidence in people situations, before and after, by survey

What does a first proof look like?

Invite 30 to 50 managers across two functions to use the coach in Slack or Teams for six weeks, with the privacy model stated up front. Nobody in HR reads a conversation.

You'd call it working when

Managers come back unprompted, rate the plans as usable, and the aggregate topic view tells L&D something they did not know.

What usually goes wrong?

  • Announcing it as a tool HR rolled out; managers assume HR reads it and stay away
  • Coaching that stays generic because company values and policies were never loaded
  • Asking for HRIS or performance access to personalise; it breaks the privacy model

What are the guardrails?

  • Conversations private to the manager; never shown to their manager, HRBP or HR
  • Not an HR record; transcripts not retained for HR purposes and never used in any process about anyone
  • Only aggregate capability and topic trends leave the system, above minimum group size
  • Reads no HRIS, performance or pay data; knows only what the manager types
  • Points to HR, legal or an ethics line when a situation needs a formal process

What leaves your boundary is set per build; the inputs above are the ceiling, and where the model runs, what it retains, and the DPA are agreed with your security team before anything is connected.

Our read

Proven sensitivity medium Order: a first build

Widely deployed and well understood. Low-risk to build well.

Some ATS, HRIS or LMS suites ship a version of this. Where yours already does the job well, switch it on. The agent earns its place when the native feature is missing, rigid, or does not respect your rules; the strategy month is where that call gets made.

Managers describe real situations involving named people, so conversations must be private to the manager, kept out of HR records, and reported to HR only as aggregate capability trends

Where it sits in the order

Needs only chat access and an agreed privacy model; it is where managers decide whether AI in this area can be trusted

We built and ran this one

This is the agent Culturro built as Merlin, an AI coach for managers, and has run inside enterprises for more than two years. Ask us anything about it.

Read the Merlin story

Is a Manager Coaching Agent worth building for your function?

That depends on your numbers, your data, and what else is on the map for you. The strategy month works that out.

How the strategy month works

Book a call

Thirty minutes. Bring the number this would move.