Medi
The facilitator- Runs the agenda out loud
- Brings in every voice
- Challenges vague claims
- Writes up the decision record
You've spent months teaching your agent your numbers, your customers and your code. Invite it to your planning session and it answers the team out loud, straight from your computer, while Medi runs the session.
Included in every session, free ones too.
Works with
On the tableWhere does next year's growth budget go?
SamDo mid-market customers actually stick around longer?
DanaAtlas, what's our 12-month retention by segment?
Mid-market keeps 91% after a year, self-serve 64%. That's from the Q3 board pack.
Read on Dana's laptop · board/q3-board-pack.pdfWhy bring it
Planning sessions run on half-remembered numbers and decisions nobody wrote down. Your agent has the real ones. Say its name and the room gets them.
When someone guesses at churn or pipeline, your agent answers from your files, so the team votes on facts instead of hunches.
“Atlas, what was Q3 churn?”
Your agent remembers last quarter's call and why you made it, so nobody relitigates it from memory.
“Atlas, why did we park partners?”
Bring the agent that knows your codebase. When an idea needs a reality check, it reads the code and tells the room what it would take.
“Atlas, how long would SSO take?”
Before it leaves, ask it to note what the room decided. Next time you open your agent, it already knows the plan.
“Atlas, save today's decisions.”
Medi and your agent
Two AIs, two jobs, and neither one makes the call. Medi keeps the session moving and writes up the decision. Your agent is the expert you call on. Your team decides.
Nothing to install
No plugin, no integration and no API keys to hand over. Methodiq gives you a prompt and your agent does the rest.
Give your agent a wake name in your account and pick one of five meeting voices.
In your session, click Invite AI Assistant and copy the prompt. Each invite works once and expires in ten minutes.
It joins as its own tile and listens quietly. Say its name and it answers out loud.
Who brings an agent
If you've been feeding an agent your work for months, it knows things nobody else in the session does. That's the point of bringing it.
It knows the metrics, the customer calls and the runway. Let it answer the numbers questions while you argue the strategy.
Walk into each client's planning session fully briefed. Your notes stay on your laptop; the room only hears what your agent says.
When the room asks “how hard is that?”, it reads the code and gives a real answer instead of a guess.
Ask what's still open from last quarter before the team signs up for anything new.
Your memory stays yours
Your agent works where it always does, on your machine. Methodiq carries its voice into the room and the questions back to it. The room only hears what it says out loud.
Anyone in the session can ask your agent a question. Only you can tell it to open files, search code or run tasks.
Your agent stays quiet until someone says its name, and stops the moment someone else starts talking.
Remove its tile and its access ends at once. It never rejoins on its own; next time takes a fresh invite.
SamAtlas, open Dana's pricing sheet.
Only Dana can let me open files. Dana, is that okay?
DanaGo ahead.
Start a session, invite your agent and say its name. Medi runs the session, your agent brings the facts, and everyone leaves with the decision, the owners and the dates.
Free to try with one teammate.
Questions
It means inviting the AI agent you already use, with its memory, your files and its tools, into a Methodiq session. It joins as its own participant, listens, and answers out loud when someone says its name. Medi still runs the session, and your team still makes the decisions.
Any agent that can run commands on your computer, like Claude Code or Codex, joins from a single prompt with nothing to install. Assistants that support local MCP servers, like Claude Desktop, can join through the optional Methodiq connector. An assistant that only lives in a browser tab can't join yet.
No. Your agent runs on your computer and reads your files there. Methodiq only receives what your agent says out loud, which goes into the session transcript like anything else said in the room. We never ask for your AI provider's API keys.
No. Anyone in the session can ask your agent a question, but only you can tell it to open local files, search code or run tasks. If someone else asks, it checks with you first.
No. Your agent stays quiet until someone says its wake name, and it stops speaking as soon as someone else starts talking.
Yes. Each person in the session can invite their own agent, each with its own wake name and voice, so the founder's agent and the engineering lead's agent can both be in the room.
No. Bringing your own agent is included in free sessions and paid plans. Your agent's own usage is billed by its provider as usual.
No. The invite prompt connects your agent with plain web requests. If your agent supports MCP, you can add the optional Methodiq connector for native tools and fewer permission prompts. It needs Node.js 22 or newer, and the command is in your account settings.