What is an L10 meeting and why does it need preparation?

An L10 meeting is the weekly leadership meeting prescribed by the Entrepreneurial Operating System (EOS). It follows a strict 90-minute agenda: segue, scorecard review, rock review, customer/employee headlines, to-do list, IDS (Identify, Discuss, Solve), and conclude. Preparation matters because without pre-compiled data on scorecards and rocks, the team wastes meeting time pulling up numbers and debating what is on track versus off track. A well-prepared L10 brief lets the team spend their 90 minutes solving problems instead of gathering information.

Can an AI agent replace the L10 meeting facilitator?

No. The facilitator role in an L10 meeting requires real-time judgment — reading the room, keeping discussions on track, calling out tangents, and deciding when to move an issue to IDS. An AI agent handles the data preparation and delivery layer: compiling scorecard numbers, flagging off-track rocks, surfacing issues from project tools, and formatting everything into a structured brief. The facilitator still runs the meeting. The agent just makes sure they walk in fully informed.

How does the AI agent pull data from Asana for L10 prep?

The agent connects to Asana through its built-in project tool integration. Once connected, it can read tasks, projects, custom fields, due dates, and completion status across all teams and projects in your Asana workspace. For L10 prep, you configure the agent to query specific projects or portfolios that map to your rocks and scorecards, then format the results into the sections of your L10 agenda. This runs on a cron (scheduled task) schedule so the brief is ready before your meeting.

How long does it take to set up automated L10 meeting briefs?

Most EOS companies can configure a working L10 brief within one to two hours. The bulk of the time is mapping your rocks and scorecard measurables to specific Asana projects, tasks, or custom fields. Once that mapping is defined in your agent's HEARTBEAT.md file, the cron runs automatically every week. Refinements happen over the first two or three meetings as you adjust which data points to include and how to format them.

What tools does the AI agent need access to for L10 prep?

At minimum, the agent needs access to your project management tool (Asana, Linear, or similar) where rocks and to-dos are tracked. For scorecard data, it depends on where your measurables live — if they are in Asana custom fields, the same connection covers it. If they are in spreadsheets or other systems, you may need additional integrations. For delivery, the agent sends the finished brief via Telegram, Slack, or email. ClawAgora agents support all three channels out of the box.

How to Use an AI Agent for EOS L10 Meeting Prep (Rocks, Scorecards, Issues)

The Monday Morning Problem Every EOS Company Knows

You run on EOS. Every Monday (or Tuesday, or whatever day your leadership team chose), you sit down for your L10 meeting. The agenda is the same every week: segue, scorecard, rock review, headlines, to-do list, IDS, conclude. The structure works. That is why you adopted it.

But here is what actually happens in a lot of L10 meetings: the first 15 to 20 minutes are spent gathering information that should have been gathered before the meeting started. Someone pulls up the scorecard spreadsheet. Someone else checks Asana for rock status. A third person tries to remember which to-dos from last week got done. By the time the team gets to IDS -- the part where real problems get solved -- the clock is already running short.

The L10 format is designed to be tight. Ninety minutes, no more. Every minute spent on data gathering is a minute stolen from problem solving. And the problem compounds: when people know the meeting will start slow, they prepare less. When they prepare less, the meeting starts slower. The cycle feeds itself.

This guide shows you how to break that cycle using an AI agent that does the preparation work automatically, on a schedule, every single week.

What a Good L10 Prep Document Looks Like

Before we get into the technical setup, let us define the output. A good L10 prep brief covers every section of the meeting agenda with the data the team needs to discuss it. Here is what that looks like:

The Sections

L10 Segment	What the Prep Brief Should Include
Segue	Nothing needed from the agent -- this is personal good news shared live
Scorecard	Each measurable with its current number, target, and whether it is on or off track for the trailing 13 weeks
Rock Review	Each rock with its status (on track, off track, done), percent complete, and any blocked dependencies
Customer/Employee Headlines	Flagged items from project tools or communication channels worth discussing
To-Do List	Carry-forward to-dos from last week with completion status
IDS	Issues surfaced from off-track rocks, missed scorecard targets, or flagged headlines
Conclude	Recap template (filled in after the meeting, not before)

The agent generates sections one through six before the meeting. Section seven is filled in live.

Step 1: Map Your EOS Structure to Your Project Tool

The first thing the agent needs is a clear mapping between your EOS structure and where the data lives. If you use Asana (which is common for EOS companies managing multiple teams), this means identifying:

Rocks: Which Asana projects or sections represent your quarterly rocks? Many EOS companies create a "Q2 2026 Rocks" project with one task per rock, assigned to the rock owner, with a custom field for status (on track / off track / done).

Scorecard measurables: Where do the weekly numbers live? Some companies track them as Asana custom fields on recurring tasks. Others use a dedicated scorecard project with one task per measurable, updated weekly. If your scorecard lives in a spreadsheet, you will need to either migrate the key numbers into Asana or set up a separate data pull.

To-dos: These are typically tasks in a dedicated "L10 To-Dos" project or section, created during each meeting and due the following week.

Issues list: A running list of unresolved issues, usually maintained as tasks in an "Issues" project or as a section within your L10 project.

Write this mapping down. You will use it in Step 3 when configuring the agent's scheduled task.

Step 2: Connect Your Agent to Asana

On ClawAgora, your AI agent runs as a managed instance with persistent access to the tools you connect. Asana is one of the supported project tool integrations.

Once connected, the agent can read across your entire Asana workspace -- all teams, all projects, all tasks. For an advertising agency running a dozen-plus teams and scores of projects, this is exactly the kind of scale where manual L10 prep becomes painful and automated prep becomes essential.

The connection is configured once during your agent's initial setup. If you have already connected Asana for other purposes (daily briefs, task tracking), you do not need to connect it again. The same connection serves all scheduled tasks.

For details on connecting project tools and communication channels to your agent, see the guide on setting up an AI chief of staff.

Step 3: Configure the L10 Brief in HEARTBEAT.md

HEARTBEAT.md is the file in your agent's workspace that defines scheduled tasks -- crons that run automatically at times you specify. Think of it as your agent's calendar of recurring work.

Here is what an L10 prep cron entry looks like conceptually:

The Schedule

For a Monday 9:00 AM L10 meeting, you might schedule the brief to generate at 7:00 AM Monday morning. This gives the team two hours to review it before sitting down. Some companies prefer Sunday evening delivery so leadership can glance at it over coffee Monday morning. Either works.

The Instructions

The cron entry tells the agent what to do when the schedule triggers. For an L10 brief, the instructions cover:

Pull scorecard data. Query each measurable from its Asana source. Compare the current value against the target. Flag any measurable that has been off track for two or more consecutive weeks.
Pull rock status. Query each rock task. Read the status custom field. Calculate implied percent complete based on subtask completion if applicable. Flag any rock marked off track or any rock with overdue subtasks.
Pull to-do completion. Query the L10 to-do list from last week. Mark each as done or not done. Calculate the completion percentage (the EOS benchmark is 90 percent or higher).
Surface issues. Compile a list of: off-track scorecard items, off-track rocks, incomplete to-dos, and any tasks across the workspace that have been flagged or escalated during the past week.
Format the brief. Assemble everything into the L10 agenda structure: scorecard table, rock review table, to-do list with checkmarks, and issues list prioritized by severity.
Deliver. Send the formatted brief to the designated channel -- Telegram group, Slack channel, or email distribution list.

What the Output Looks Like

The delivered brief reads like a structured document, not a wall of text. Here is a simplified example of what the scorecard section might contain:

Measurable	Owner	This Week	Target	Status	Trend (4 wk)
Revenue booked	VP Sales	$47,200	$50,000	Off track	Declining
New leads	Marketing Dir	38	30	On track	Stable
Client NPS	Account Lead	72	75	Off track	Improving
Billable utilization	Ops Dir	81%	85%	Off track	Stable

The rock review section follows a similar table format, and the issues section is a prioritized bullet list with context pulled from the underlying tasks.

Step 4: Refine Over Two or Three Meetings

No L10 brief is perfect on the first run. The most common adjustments after the first meeting:

Too much data. The agent pulled every task and every number. You realize you only need the top-level measurables and the quarterly rocks, not every subtask. Narrow the scope in the cron instructions.

Missing context. A rock shows as "on track" in Asana but the owner knows it is actually behind. This is a data hygiene issue, not an agent issue -- the agent can only report what is in the tool. Use this as motivation to keep Asana updated, which is itself a benefit of the system.

Wrong delivery time. 7:00 AM was too early or too late. Adjust the cron schedule. Some teams land on Friday afternoon delivery so leadership has the weekend to review if they want to.

Format preferences. Maybe the team wants the issues section first instead of last, because that is what they actually discuss the most. Reorder the format instructions in HEARTBEAT.md.

By the third meeting, the brief should feel like it was written by an experienced executive assistant who sat in on every meeting and tracked every number.

Why This Matters More Than It Seems

The direct benefit is obvious: the team saves 15 to 20 minutes per meeting and walks in better prepared. Over a year of weekly L10s, that is 13 to 17 hours of reclaimed leadership time. For a company with six people in the room, that is 78 to 102 person-hours.

But the indirect benefit is larger. When the L10 brief arrives every Monday morning, fully formatted and impossible to ignore, it creates accountability pressure. Everyone can see which rocks are off track, which scorecard numbers are missed, and which to-dos did not get done. The data is there in black and white before anyone sits down.

This is the dynamic that EOS is designed to create. The system works when the data flows. An AI agent that automates the data flow makes the system work better.

Extending Beyond the Weekly L10

Once the weekly L10 brief is running, the same pattern extends to other EOS rhythms:

Quarterly planning prep. Before each quarterly planning session, the agent can compile a comprehensive review of all rocks from the past quarter, scorecard trends over 13 weeks, and a draft issues list for the upcoming quarter.

Annual planning data. The agent can generate year-in-review summaries of scorecard performance, rock completion rates, and issue resolution patterns to inform the annual planning session.

Daily standups. If your team runs daily huddles in addition to weekly L10s, a lighter version of the same cron can deliver a daily brief covering open to-dos and any items that need immediate attention.

For more on configuring daily scheduled tasks, see the guide on scheduled tasks and daily routines.

What the Agent Does Not Do

It is worth being explicit about boundaries. The AI agent handles the data layer of L10 preparation. It does not:

Facilitate the meeting. You still need a human facilitator who keeps the discussion on track and manages the clock.
Make judgment calls about priorities. The agent surfaces issues. The leadership team decides which ones to IDS.
Replace accountability. If a rock owner is not updating their status in Asana, the agent will report stale data. The discipline of updating the tool is still a human responsibility.
Handle the segue. Personal and professional good news is shared live. That is the point.

The agent is the preparation engine. The humans run the meeting.

Getting Started

If you are running EOS and your leadership team spends the first chunk of every L10 gathering data instead of solving problems, an AI agent configured with cron-driven L10 briefs can fix that in a week.

The setup takes one to two hours. The mapping of your EOS structure to Asana is the most important step. Once that is done, the agent handles the rest every week, automatically, without being asked.

ClawAgora plans that support scheduled tasks start at $29.90 per month on Spark. For companies running the full EOS framework with multiple teams, the Blaze plan at $1,090 per year provides the capacity for complex, multi-source briefs.

Start with the weekly L10 brief. Once you see it working, extend to quarterly planning prep and daily standups. The pattern is the same -- define the data source, define the format, set the schedule, choose the delivery channel.

Your L10 meetings were designed to be the most productive 90 minutes of your week. Let the agent handle the prep so your team can spend every one of those minutes on what matters.

For a full story of how a 20-person agency set this up in three days, read How a 20-Person Agency Replaced Their Departing Operations Director with an AI Agent. ClawAgora plans start at $29.90/month with managed hosting and AI credits included. See pricing and get started.