As a CEO, your mornings set the tone for everything that follows. Mine used to start the same way every day: open Gmail, scan for fires. Switch to Slack, scroll through channels and DMs. Check Google Calendar. Open Jira. Glance at Notion. Pull up Granola to review yesterday’s meeting notes. By the time I’d pieced together a picture of what actually needed my attention, forty-five minutes had evaporated—and I hadn’t done a single productive thing yet.
That changed about a month ago. Now, before I finish my first coffee, I get a single prioritized briefing that tells me exactly what matters today—sourced from every tool my team uses, synthesized by an AI agent I built in under an hour.
Here’s how I did it, and why I think every founder running a scaling company should consider doing the same.
The Problem Isn’t Information—It’s Attention
Most founders I know aren’t short on information. We’re drowning in it. The average SaaS company uses around 100 tools, and even at a lean startup, the number of surfaces where important things land is staggering. A critical action item gets buried in a Slack thread. A commitment you made in a meeting two days ago sits in a transcript nobody re-reads. A deadline for an event registration is quietly ticking down in your inbox while you’re heads-down on something else.
The cost isn’t missing one thing. It’s the cognitive overhead of checking everywhere, all the time, to make sure you haven’t missed something. That vigilance tax is real, and it compounds. It makes you reactive instead of strategic. It turns your first hour of the day—the hour when your thinking is sharpest—into a scavenger hunt.
I realized I didn’t need another dashboard or another notification channel. I needed something that could do the scavenger hunt for me, every single day, before I even sat down.
What I Built
I built a scheduled AI agent using Anthropic’s Cowork tool—a desktop companion that can connect to the services you already use and run tasks on a schedule. The setup took less than an hour. No code. No developer required. Just a clear description of what I wanted and the right integrations connected.
The agent runs twice a day. At 8 AM, it delivers a morning action briefing. At 1 PM, it runs again with an afternoon check-in—a course correction for the second half of the day. Each run pulls from five sources:
-
Gmail (unread and important messages from the last 48 hours),
-
Slack (DMs, mentions, and key channels),
-
Google Calendar (today’s meetings and all-day events),
-
Granola (action items and commitments from recent meeting transcripts), and
-
Notion (task databases and strategic documents).
It then synthesizes everything into a single briefing organized by urgency: red items that are time-sensitive or overdue, yellow items that are important but not on fire, and green items to keep on my radar. Every item includes the original source link so I can act on it immediately.
The Design Choices That Matter
Twice a day, not real-time
I deliberately chose not to build this as a real-time notification system. Notifications are part of the problem—they interrupt deep work and train you to be reactive. A scheduled briefing respects your attention. It says: here’s what matters, in one place, when you’re ready to look at it. The morning run sets the plan. The afternoon run adjusts it.
Smart triage, not just aggregation
The agent doesn’t just dump every unread message on my desk. It evaluates urgency based on context: explicit deadlines in emails, promises I made in meetings (“I’ll send that tomorrow”), direct asks from team members, SLA approaching their due dates. It understands the difference between “nice to know” and “you said you’d do this and you haven’t yet.”
Draft, but never send
This was a non-negotiable principle. The agent can draft suggested replies in Slack and Gmail, but it never sends anything on my behalf. Every action requires my review and approval. AI should reduce friction, not remove accountability. My team hears from me, not from a bot pretending to be me.
What Changed
The quantitative stuff is easy to point to. I reclaimed roughly 45 minutes every morning—time I used to spend cycling through apps trying to assemble a mental model of my day. Response times on things that actually matter improved because I’m seeing them surfaced by priority rather than stumbling across them.
But the qualitative shift is what surprised me most.
Last Saturday, the briefing flagged an event registration deadline that was closing the following evening. It was buried three emails deep in a thread I hadn’t opened. Without the agent, I would have missed it entirely—it was a founder forum I’d been personally invited to.
The same week, it surfaced a commitment I’d made during a partner meeting to send an email outlining four priority items. Three days had passed. The meeting transcript was sitting in Granola, and the action item was clear in the notes, but I hadn’t gone back to look. The agent caught it and put it at the top of my briefing.
It also flagged that I’d told a colleague “I’ll get that set up tomorrow” in a Slack DM—and tomorrow had come and gone. Small commitments like that are the ones that erode trust when they slip, precisely because they feel too minor to track formally.
The Biggest Win Wasn’t Time—It Was Clarity
The most valuable thing this agent gives me isn’t the 45 minutes back. It’s the confidence that nothing is falling through the cracks.
That might sound like a small thing, but if you’ve run a company through a growth phase, you know the feeling: a low-grade background anxiety that something important is sitting in a channel you haven’t checked, or a follow-up you promised is quietly aging. That anxiety doesn’t just waste time—it degrades the quality of every conversation you’re in, because part of your brain is always scanning instead of being present.
When you know the scan has already happened—thoroughly, across every surface, before your day started—you show up differently. Meetings are better. Decisions come faster. You’re working from a place of clarity instead of vigilance.
The Afternoon Surprise
I almost didn’t build the afternoon run. It felt redundant—why would I need a second briefing? But it turned out to be the more valuable of the two.
By 1 PM, your morning plan has collided with reality. Meetings ran long. A new Slack thread pulled you into something urgent. The afternoon briefing recalibrates. It catches anything new that came in, resurfaces morning items you haven’t acted on yet, and gives you a clean priority list for the remaining hours. It’s a reset button for the second half of your day.
You Can Build This Today
I want to be clear about what this required: no engineering team, no custom API integrations, no code. I described what I wanted in plain English, connected the services I already use, and set a schedule. The whole thing was running within an hour.
The tools to do this exist right now. Anthropic’s Cowork, which I used, is one option. The specifics matter less than the principle: if you’re a founder spending your first hour of the day just getting oriented—cycling through apps, triaging inboxes, trying to remember what you promised in yesterday’s meetings—that hour is recoverable.
AI is at its best not when it’s replacing your judgment, but when it’s clearing the path so you can use it. Build the agent. Get your mornings back. You’ll be surprised how much changes when you start every day already knowing what matters.
Chris Grimes is the CEO & Co-Founder of FundMore, a mortgage technology platform helping lenders modernize their operations across North America.