8:42 AM - Inbox Triage
I pour my second coffee and open the Solar AI dashboard. The morning metrics are waiting. I can see in the top card that 847 emails went out across the network overnight. The open rate sits at 28%, which is solid. Four demos got booked yesterday through the automated scheduling link, which means four new prospects are on the calendar for this week. Revenue from recurring customers hits the Stripe dashboard at $3,412 for the month so far. We're tracking to beat last month.
I flip to Gmail and start reviewing the outbound emails the system drafted yesterday. They're sitting in a "Pending Approval" label. The first one is addressed to Marcus Chen at Sunbow Solar in Arizona. The body is good. The system picked up that Marcus's company installed 847 systems last year from their public LinkedIn and referenced their recent partnership with a local contractor. The email doesn't feel generic. I hit approve.
The next three are equally solid. The fifth one, though, has a problem. It's addressed to "Dear Solar Decision-Maker" because the prospect list didn't include a contact name. The system should have rejected this one outright. I flag it in the rules file - no sending to generic titles. I've been running this for six weeks now and it's the third time this month. I add a note to the engineering backlog that the name-validation check needs teeth.
I turn it down and add a manual note to the draft: hit the prospect's LinkedIn, find the actual owner, try again tomorrow.
I check the Slack alert log. Three customer notifications came in overnight. One is from Tracy Valdez at Verde Solar asking about monthly reporting. One is from a company called Apex Energy asking why they haven't seen new leads yet (they signed up ten days ago). The third is a churn notification: Colton's Solar Solutions turned off billing. That stings. They were in month four.
10:15 AM - The Churn Call
I pull up Colton's Solar Solutions in the admin panel. I can see their full history. They came in through a LinkedIn message. First month they got 23 meetings booked. Second month was 16. By month three it dropped to 8. The system kept sending, but something shifted. I look at the conversion rate chart - their meeting-to-close ratio is 3.2%, which is actually above our average of 2.1%. So it wasn't quality. They just stopped booking.
I call Colton directly. His number is in the intake form.
He picks up on the second ring. "Hey, yeah, we just couldn't keep up. We got swamped with jobs and didn't have the capacity to follow up on the leads you guys were sending. It's not on you."
I ask if we can pause the subscription instead of canceling. He says he'll think about it. I offer to pause it for 30 days at no charge, and if they restart it, I'll hand-adjust their daily volume from the usual 30 emails a day down to 15 while they clear their backlog.
He takes it. I pause the account, add a note to follow up on day 28, and move on.
The pause isn't ideal, but it's better than a churn. I'll probably end up losing them eventually - a company that size either does the work or they don't. But the work I did kept the door open.
12:30 PM - Lunch and the Metrics Check
I grab lunch at my desk and pull up the full weekly dashboard in a Google Sheets file I've been building. It's not fancy, but it's mine.
Week to date: 5,891 emails sent, 1,651 opens at 28%, 487 clicks on demos at 8.3% of opens, 17 demos booked, 4 customers converted to paid subscriptions. Revenue this week is $8,240. Last week was similar. The month is on pace to hit $12,500 in new recurring revenue and sit around $47,000 in total MRR.
But there's a soft spot. The demos are still there, but the close rate is dropping. It was 24% last month. This month it's tracking at 19%. I need to watch that. My customer-success motion is just me sending follow-up emails three days after each demo. Maybe I'm losing them in week two when they haven't decided yet.
I make a note to record a short video walk-through of Solar AI and send it to the next batch of demo attendees in their follow-up sequence. Might be enough to push them over the line.
2:08 PM - The Billing Question
Tracy Valdez from Verde Solar sent a detailed email about their monthly reporting. They want custom dashboards and real-time prospect status tracking. That's not a feature I have. The system shows which emails were sent and which got opened and clicked. That's as detailed as it goes right now.
I write back. I tell her what we track and what we don't. I offer to set up a weekly call where I manually pull a custom report for them - at least for the next 30 days while we see if that solves the problem. If it works, I'll consider building it into the product properly.
She replies within the hour saying that would be perfect.
This is the work that doesn't scale. But it keeps customers happy. I add her to my Tuesday calendar for next week.
4:30 PM - Pipeline Review
I open the CRM view, a simple Airtable base that mirrors what Solar AI tracks. Current pipeline is 127 prospects. 19 are in "Demo scheduled this week." 31 have opened emails but haven't clicked the scheduling link yet. 77 are still in the first-touch sequence.
Of the 19 demos, I spot one that looks promising. A company called Brightsun Energy in Nevada. They've been contacted twice, opened both emails, and clicked the link yesterday. The demo is scheduled for Thursday at 3pm.
I add a note to my Thursday morning calendar: prepare a demo tailored to their company size and service area. I'll pull their LinkedIn, look at their sales team size, their recent hires. I'll customize the first 90 seconds of the demo instead of running the standard opener.
The other numbers look normal. Conversion from "email sent" to "demo booked" is about 3%. The close rate from demo to paid is still the problem.
I send a message to the Slack channel I use to talk to myself. Note to self: A/B test the follow-up sequence. Maybe people need a softer ask. Maybe the pricing page is turning them off. Maybe they need a second conversation, not just email.
5:47 PM - A Clean Convert
I see a notification that a prospect from last week just converted. Her name is Priya Kapoor. She owns a small solar installation company in California. She attended a demo eight days ago, got the follow-up email on day three, and just filled out the conversion form. Her payment went through Stripe. She's set up for her first month of service starting tomorrow.
I do something I told myself I would do whenever this happens: I send her a direct email. Not automated. From me. I write:
"Hey Priya, thank you for signing up with Solar AI. I built this because I spent five years watching solar companies work harder than they should to fill their pipeline. You're going to get good leads next week. If any of them don't feel right or don't fit your service area, just tell me and I'll adjust. Excited to work with you."
It takes three minutes. But she'll get it tomorrow morning from my personal email, and it sets a different tone than the automation.
6:15 PM - Wrap
I close the laptop. The day is good. Four customers happy, one churn paused, one new conversion, one small engineering bug flagged, and the pipeline is clean.
This is not passive income. It's not a business where I set it and forget it. Solar AI does the heavy lifting. It writes and sends 6,000 emails, tracks opens, schedules meetings, processes payments. But I'm still doing the real work. I approve what goes out. I talk to customers who are stuck. I fix things that break. I think about why conversions are dropping and run the tests. I customize demos. I handhold the companies that need it.
The machine amplifies me. It doesn't replace me. And at six weeks in, that feels fair. The revenue is real. The metrics are on track. If the close rate keeps dropping, I'll need to either get better at the follow-up or build more into the product. For now, I have the bandwidth to do both. Tomorrow I'll start the A/B test.