Skip to main content

​
Video walkthrough

Exec MCP connects your workspace to the AI assistant you already work in, alongside whatever else you have connected there. Once it is in place you ask questions in plain language and get answers built from live Exec data. There are no reports to run and no need to remember which dashboard a number lives on. For a team lead that changes the shape of the work. Instead of exporting data, reading it, and deciding what to do about it as three separate jobs, you do all three in one conversation, then act on the answer without leaving the tool you were already in. Most of what follows needs workspace admin or group admin, which is the same permission that governs analytics inside Exec.

​
1. Start with a readout

What comes back is a table per group with headcount, sessions per person, and the gold, silver, and bronze split, followed by a written read of it.
Exec hands your assistant one group at a time rather than a ready-made comparison table, so the assistant pulls each group and assembles the breakout itself. Two consequences worth knowing: your groups have to exist in Exec for this to work at all, and naming the two or three groups you actually care about is quicker than asking it to do every group in the workspace.
Three things are worth looking for in the write-up:
  • The scenarios dragging the average down. Usually two or three, not a general malaise. Those are what you build against.
  • Where practice time is actually going. This is the finding people do not expect. A team can be practicing plenty and still be spending most of its minutes on scenarios that have nothing to do with what they are paid to do. Ask what share of practice minutes went to selling scenarios? if the readout does not volunteer it.
  • The biggest improvement, not just the lowest score. Improvement from first attempt to best attempt is the number worth taking into a team meeting.
If your assistant renders panels, ask for it as a dashboard: give me that as a dashboard.

​
2. Find who needs coaching

You get two shortlists with names and groups, and, more useful, the reason each person is on the list. The list of strugglers is the one that turns into work. It is a coaching list: the people to put on your one-on-one agenda this week, with a specific thing to talk about rather than a general sense that they are behind. Go one level deeper on anyone you are actually going to sit down with:
Quoting the actual moment lands far better in a coaching conversation than repeating a summary, and it lets you judge for yourself whether the grade was fair.
Superlative questions like “who struggled the most” pull the numerically worst row, which on a real workspace is often somebody who opened a session and abandoned it. Nobody performed badly there. Nobody performed. Add ranked to the question, and read Ask for the real thing before you take any of these names into a review.

​
3. Cross-check against real calls

If you have call scoring set up, and objection handling is one of the things your scorecards capture, the same conversation can ask what is happening on live calls, not just in practice.
You get the recurring themes, how often each one was handled, and the calls where they came up. The pattern is usually blunter than expected: a particular objection that gets folded on every single time it is raised, or one that never gets engaged with at all. That is the strongest input you can give the next step, because it is evidence from real conversations rather than from practice.

​
4. Build the practice that closes the gap

The payoff. Because everything above is in the same conversation, you do not have to restate any of it.
Your assistant will usually come back with a couple of clarifying questions and a view on which gap to aim at. Answer them, and it writes the scenario into your drafts in Exec. From there you can open the draft in Exec and keep working on it, or stay where you are:
Then assign it, and you have gone from a question to a piece of practice aimed at a real weakness without exporting anything.

​
Set up a project so you do not repeat yourself

This is the habit that makes the difference between trying MCP once and using it every week. Most assistants let you save a workspace with its own standing instructions. Claude calls these projects, ChatGPT has custom instructions, and several other tools have their own version. Put Exec in one, and write down the things you would otherwise retype every time:
  • Who your teams are and how you want them broken out
  • The house rules for numbers, such as always saying what a claim is based on, and only counting ranked sessions
  • Your usual format, if you always want a short written summary rather than a table
  • What to leave out, which matters if you are going to share your screen
After that, “give me the weekly readout” is enough, because the standing instructions carry the rest. It is the same reason a saved report beats rebuilding a filter every Monday.

​
Make it your own

Everything above is a starting point rather than a script. The prompts that end up mattering are the ones you invent for your own week, so it is worth spending a few minutes asking for things this page does not mention and seeing how far it goes. If something does not work the way you expect, Troubleshooting covers the failures people actually hit, and [email protected] will pick up anything it does not.

​
Next

Prompts and reliable answers

The full prompt library, and how to check an answer before you act on it

Managing the org

The same walkthrough one level up, across every group

Admin controls

What you can govern before you roll this out