8:45 AM - The Dashboard
I open my laptop with a coffee that's still hot. The dashboard loads. Six new signups overnight from the SEO content about lost deals. Seven emails in the main inbox. Three Slack alerts from the product dashboard flagged with a orange warning. Week to date, we're at $8,800 in MRR. That's good. On pace for month-end.
I scan the Slack channel where the outreach agent posts its daily summary. It's prepared fifteen cold emails to outbound prospects and wants approval before they go out. I'll get to those in twenty minutes. First, I need to see what actually happened last night and this morning.
One of the three Slack alerts is about a customer churn. AccountLink Solutions, a small consulting firm that signed up six weeks ago, just cancelled their subscription in Stripe. I click into their account history. They signed up, went through onboarding, ran two win-loss analyses on their own. Then nothing for three weeks. Now gone. The churn message includes a note: "too much manual work to load our past sales data." I add a note to the churn folder: ask this question in the next batch of cancellation surveys. This is the first churn I've seen, and I need to understand if it's a product issue or a ICP issue.
The other two Slack alerts are false positives. Billing system caught a duplicate row in the customer table. Database guy didn't clean up right. I flag it for later. The third alert is just a reminder that two customers are coming due for renewal this week. I make a note to email them Friday.
9:20 AM - The Outreach Queue
I pull up the draft email thread in Gmail. The agent has prepared a sequence for Marcus Webb, VP of Sales at CleanOps, a 280-person manufacturing software company. The first email subject is "Why your best deals fell through at the close."
I read through it. It's good. It hooks on a specific pain point, mentions a statistic about decision drivers, and offers a brief demo call. The agent has pulled Marcus's name from LinkedIn, referenced a recent company news article about their Series B, and kept the tone conversational. Not slick. Not spammy. I approve it.
I look at the next five emails in the queue. Two of them are too long. One has an awkward opening line. I send a note back to the agent in Slack: "Emails 3-4 are too direct on the ask. Pull back 20%. Soften the CTA on 7. Reference their website case study like I did in the Marcus one." The agent will regenerate those three by end of day.
I approve the other ten and they go out. That leaves eight still in draft. I don't have time for all of them right now.
10:40 AM - The Flagged Conflict
My phone buzzes. A Slack DM from Sara Nakamura, one of our three paying customers. She's the ops lead at a SaaS company called RentFlow. Her message: "Can I get a report on why we lost the Midland Bank deal? The output your system generated seems off."
I open her account in our dashboard. RentFlow ran a win-loss analysis on their lost deal last Thursday. The system interviewed four people from their side. Midland Bank said no. The analysis came back with three decision drivers: 1. Pricing, 2. Integration roadmap timing, 3. Sales process fit.
Sara says the output is "off." I call her to understand what she means. She explains: her team thinks pricing was a factor, yes, but the real issue was that Midland's CTO wanted custom integrations that RentFlow's product roadmap didn't support. The analysis flagged "integration roadmap timing" but not the actual scope of what they needed. In the raw interview notes, one comment says "they wanted SDK-level customization and we said no."
This is a real problem. The AI is capturing the theme but not the actual scope. I take a note. I tell Sara I'll regenerate the analysis with a focus on the scope question. She's satisfied. I add this to a list in Linear for the product team: "Analysis sometimes misses scope of customization requests in integration conflicts."
12:15 PM - Metrics and a Win
I stop for lunch at my desk. A bowl of rice with grilled chicken. I open Stripe to check on MRR and churn.
Current MRR: $9,100 (we're picking up a new customer today, looks like). Week-to-date signups: 18 (we're tracking ahead of the 20-per-month goal). One churn so far (AccountLink). Four renewals coming in the next ten days. One customer is two days late on a payment. I shoot a gentle email: "Hey, didn't see the payment for April. Want to update the card on file?"
Then an email arrives from Nicole Soares at Vertex Solutions. I know this name. Vertex was a demo two weeks ago. No close then. I figured we'd lost them. But Nicole is writing: "We ended up using your product to analyze why we lost the Acme deal. It was exactly what we needed. We'd like to sign up for the monthly subscription."
She's even included feedback: "The report format helped us see that we lost on scope of services, not on price like we thought. That's going to change how we sell going forward."
I feel something good. This is the first time I'm seeing the product create an actual outcome for someone. Not just a tool. An outcome. I respond to Nicole immediately, thank her, and let her know I'll send over the onboarding setup within the hour. I'll monitor her account closely for the first month.
2:45 PM - The Billing Edge Case
An email comes in from Michael Torres at HiveScale. He's asking about volume discounts. He says they're planning to run 40 analyses this quarter and wants to know if we discount for bulk usage. Right now we charge $350 per customer per month. Flat rate. No volume breaks.
This is the first time anyone has asked. I don't have a policy. I could say no. I could offer something. Michael's an inbound lead from an SEO article, so I know he found us organically. That's valuable. But I can't just give away margin.
I tell him: "Let's talk through your use case this week. Forty analyses is substantial. I want to make sure the pricing works for you and for us." I schedule a call for Thursday afternoon. I add a note to Linear: "Volume discount policy needed for Q2."
4:15 PM - The Bug Fix
I notice something wrong in one of our recent customer reports. I pull the screenshot. The timeline chart is showing closed dates that are off by one day. It's a small bug but it could undermine trust if the customer doesn't catch it.
I file a ticket in Linear with a screenshot and exact reproduction steps. The product engineer responds within 20 minutes: he's found the issue and has a fix. Ship by tomorrow morning. Good. I close the ticket and thank him in Slack.
5:50 PM - Wrap
I scan my task list for tomorrow. Follow-up call with Michael Torres Thursday. Regenerate Sara's analysis. Monitor Nicole at Vertex through her onboarding. Push out the renewal reminder emails Friday. Add up the weekly numbers and prepare next week's status.
I close the laptop at 6:10 PM. Week to date: $9,100 in MRR, one churn, one surprise win, eighteen signups, three serious customer interactions, one bug caught and fixed, two emails I sent personally.
The thing that's clear now, two months in, is that this business runs on attention. The agent generates the outreach and some of the analysis. But I'm the one who decides what gets sent. I'm the one who listens when a customer says something is wrong. I'm the one who makes calls about pricing and policy. The AI is the force multiplier. But the business is me paying attention.
That's harder than I expected, and it's also why it works.