You walk into Monday’s L10 already knowing which Rocks are off track, which Issues have come up three weeks running, and who owes what from last week — and the prep that got you there took about five minutes. Sunday night stayed Sunday night, with no spreadsheet open next to your email and no scrolling back through old messages to find the thing someone raised on a call Thursday afternoon.
That’s what the four workflows below are for. Each one runs through the Strety MCP, which connects your Strety workspace to Claude, ChatGPT, or any tool that supports remote MCP servers. You ask in plain language, it reads your real data, and it writes the result back into Strety when you tell it to.
We built the connector because we wanted to run our own company from the tools already open on our screens, so everything here is something we or our customers do most weeks. The prompts are included as we use them — copy them, swap in your own names and dates, and see what comes back.
Before you start: one connection MCP is an open standard, and a well-governed one. Anthropic donated it to the Linux Foundation’s Agentic AI Foundation in December 2025 alongside co-founders Block and OpenAI, with support from Google, Microsoft, AWS and Cloudflare, and there are now more than 10,000 active public MCP servers running against it. That’s why “ChatGPT or Claude” is a genuine either/or here: you can pick whichever assistant your team already pays for, and the workflows below behave the same way in both.
Setup runs about two minutes. You open your assistant’s connector directory, add Strety, sign in and authorize, and the Strety MCP page walks through those four steps if you want them in detail. Three things are worth knowing before you start: the connection carries your existing Strety permissions, so the assistant sees exactly what you can already open and nothing beyond it, there’s no additional fee for using it, and you can connect on your own without waiting for a company-wide rollout.
If you’d like the longer explanation of what MCP is and how it differs from a traditional integration, we wrote that up separately in what MCP means for EOS teams .
One number for context before we get to the workflows. McKinsey’s 2026 State of AI survey found that nearly nine in ten organizations now use AI regularly in at least one business function, while the share of smaller organizations scaling AI agents held flat at 22 percent. Most companies our size have AI everywhere and agents nowhere, and four workflows against your own operating data is a reasonable way to close that gap — an afternoon’s work, not a project plan.
Prep an L10 in five minutes The reason L10 prep can swallow a Sunday evening is assembly rather than thinking. Your Scorecard lives in one place, the Rocks in another, last week’s To Dos somewhere else again, and the Issue someone raised mid-week is sitting in a message thread you’d have to go find. Researchers at Harvard Business Review tracked knowledge workers toggling between applications roughly 1,200 times a day, losing close to four hours a week simply reorienting themselves — and meeting prep is that tax, concentrated into one sitting.
Three prompts handle the assembly for you.
Pull last week’s Scorecard, any Rocks that are off track, all open Issues, and every To Do that came due since our last L10.
Which of those Issues have come up more than once in the last month?
Add the three worth solving to Monday’s agenda and assign owners.
What lands in Strety: agenda items and Issues with owners, ready before anyone joins the call.
The second prompt is the one that changes the meeting itself. Repeat Issues tend to be the ones your team has been working around without saying so, and they’re genuinely hard to spot when you’re reading a fresh list every week. Ask the question across a month and the pattern surfaces on its own.
Roman, founder and Visionary at Cleango, runs a daily version of the same idea. “In just one command I can understand everything that’s going on in our business,” he told us, meaning his Scorecard, his at-risk Rocks and whatever moved since yesterday, all in a single pass. (His full workflow is here , along with eight more.)
It’s still worth glancing at the agenda before you start, because the assistant orders Issues by what it read and you know things about your team that your data doesn’t. And if you’re still settling the agenda itself, our L10 agenda template and guide covers the structure.
Turn a Rock into a project workstream Every quarter seems to produce one of these: a Rock that’s been “in progress” for six weeks, owned by someone perfectly competent, with nothing visible happening behind it. Usually that’s because the Rock is a title and a due date, and the work between those two things never got written down anywhere.
Read the Q3 Rock on onboarding and tell me what’s blocking it, based on the To Dos and comments already attached to it.
Give me the workstream — the four or five steps between here and done, with an owner and a date for each. The quarter ends September 30.
Build that as a project in Strety with those To Dos assigned.
What lands in Strety: a project with real To Dos, owners and dates, attached to the Rock it belongs to.
Notice that the second prompt names the quarter’s end date, and that’s deliberate — left to its own judgment, the assistant will produce dates that sound entirely reasonable and land somewhere in October.
The same approach is the fastest way we’ve found to fix a whole set of Rocks at once. Brian, our co-founder, runs Company Rocks through the connector at the start of a quarter and asks it to flag the ones that aren’t SMART and leave feedback as comments, which turns an afternoon of opening each Rock and typing variations of the same note into one conversation.
What your Rock owner ends up with is a workstream that has their name against specific steps, and that makes for a very different exchange at the next L10 than “still working on it.”
Close out a session while it’s fresh 🎙️ The hour after a meeting is when the notes are still good and the next thing has usually already started, which is exactly why commitments made out loud at 10:40 have a way of evaporating by Friday.
Capture the commitments from this session as To Dos, with owners and due dates.
What got raised but didn’t get solved? Log those as Issues for next time.
Write me the recap in plain language, ready to send to the team.
What lands in Strety: To Dos with owners, Issues queued for the next meeting, and a recap you can paste into an email.
Eric Dykes, an EOS Implementer we work with, runs his whole post-session checklist this way. He manages his practice across Strety, a CRM and Google Workspace, so closing out a session used to mean touching all three, and now the context work happens in one conversation while the session is still fresh in his head.
The mobile version of this is better than it sounds. “The fastest way to create a To Do while on the road is just to speak into it,” Brian says — voice-to-text into the assistant on the way out of a building, no keyboard involved. Work that used to wait for a desk gets finished in the parking lot instead.
The one thing to look over is owner assignment. When a session covers people the assistant hasn’t seen much of in your workspace, it will guess, and it guesses with real confidence.
Find the doc you can’t name You remember what the doc covers, but you have no idea what someone called it eighteen months ago. This is the workflow people tend to underestimate, and it’s also the one that shows you the connection is genuinely reading your workspace.
Which of our Docs covers how we handle a refund request over $500?
What does it say about approvals?
Which Docs are getting the most comments right now?
What lands in Strety: nothing at all. This one is read-only, which makes it a safe first test right after you connect.
That third prompt turns out to be a useful signal on its own. Comment volume tends to track whatever the team is still working out, so it points you toward the processes that need a decision from you.
Shaz Khan, co-founder and Integrator at Tono Pizzeria, built himself an executive dashboard on top of this idea, and his reasoning has stuck with us: “Comprehension is an act of communication. I’m trying to reduce friction in communication.” The same underlying data, with each person able to ask for it in the shape they think in. His build is written up in the nine use cases post as well.
If you want the mechanics of Docs themselves, the complete guide lives in our help center.
Where it gets things wrong Three limits are worth knowing before you let the assistant write anything that matters.
It guesses at owners. Assignment is what it gets wrong most often, and it does so without hedging, so the owner column is the one thing to read before you accept a batch of To Dos.
It invents dates. Ask for a plan without naming your quarter’s boundaries and you’ll get deadlines that look plausible and fall outside the quarter, which is why the dates belong in the prompt.
Vague prompts produce vague To Dos. “Help me with onboarding” gets you a To Do that says roughly that, while naming the outcome you want gets you one your Rock owner can pick up.
None of these are reasons to keep the connector at arm’s length. They’re simply why you’ll want to glance over its work before it becomes your team’s week.
Frequently asked questions Can ChatGPT and Claude both connect to Strety? Yes. The Strety MCP works with Claude, ChatGPT, and any tool that supports remote MCP servers, and both the connection process and the prompts are the same in each.
Does the AI only read my data, or can it make changes? Both. It can pull your Scorecard, Rocks, Issues, To Dos and Docs, and it can write back — creating To Dos, logging Issues, checking in measurables, and building projects.
Is it safe to connect our company data to an AI assistant? The connection respects your existing Strety permissions, so the assistant can’t reach anything you couldn’t already open yourself. Our compliance certifications are published in the Trust Center.
Do I need my whole company to set this up? No. Individual team members can connect on their own, without a company-wide rollout.
Does the Strety MCP cost extra? No. There’s no additional fee for the connector on any plan.
What’s the first thing I should ask it? Ask about your own To Dos. It confirms the connection works and shows you how the assistant talks about your workspace before you let it write anything.
Pick one workflow this week 🙌 Four workflows, one connection, and the same shape each time: you ask for what you’d otherwise have assembled by hand, then let it write the result back where your team will see it.
We’d start with L10 prep, since it’s the one with a deadline attached and you’ll know within a week whether it earned its place. Run the three prompts before your next meeting and compare it against however long prep usually takes you.
The connector is included with every Strety plan at no extra cost, and it’s how we run our own company. If you’re not with us yet, you can try Strety free for 30 days — connect it to Claude or ChatGPT on day one and see what your operating data looks like when you can simply talk to it.
