8:42 AM - Inbox triage
I open Clearwater's dashboard and scan the overnight activity. Seven new inbound replies in Gmail. Three of them are the kind I like to see: "Tell me more about the pricing" and "Can you send me a demo link." Two others are automated rejections. One is a bounce.
The last one is from Carol Reyes at Reyes Family Practice in Jacksonville. It's terse: "Received two emails from you to my inbox. Not sure how, but this feels like spam." I flag that for later.
My AI agent prepared sixteen outbound emails yesterday. They're queued in drafts, waiting for my review. I pull them up in Gmail and start reading through. The subject lines are solid. "One metric physical therapists are tracking this fall" and "Your clinic's scheduling problem, solved" are better than the templated dreck I was sending before.
I spot a problem in the third email - the AI referred to a "patient appointment system" when the prospect runs a chiropractic clinic. Chiropractors call them adjustments, not appointments. It's a small thing, but it's the difference between sounding like an insider and sounding like a bot. I edit the line, change "appointments" to "adjustments," and hit approve.
By 9:04 AM, I've approved twelve emails. Four I've re-written. One I've rejected entirely because the tone felt too pushy. The Slack integration on my phone buzzes: a notification that the approved emails are now sending.
9:47 AM - The flagged conflict
I open Carol Reyes's email again and read it more carefully. She's right. There are two emails from Clearwater in her inbox. The first one I sent yesterday, personalized and targeted. The second one came through our LinkedIn integration at 2 AM - an automated connection request I didn't manually approve. I must have left the outreach setting too broad.
This is the hard part of running this business solo. The automation moves faster than my judgment. I shoot her an email directly:
"Hi Carol, I apologize. You shouldn't have received that second message. I'm the one sending these, and I manually review outreach because exactly this kind of thing shouldn't happen. I shouldn't have left an automated LinkedIn setting active without tighter targeting. I'm disabling it now. If Clearwater is useful for Reyes Family, great - if not, no worries at all. Either way, no more duplicate messages. My bad."
I turn off the LinkedIn automation in the product dashboard. Then I make a note in my Linear project: "Add reviewer approval step before LinkedIn auto-sends." I'll build that next week.
By 10:15 AM, I've logged three hours of work and haven't made a single sale. This is the reality people don't see.
11:22 AM - A conversion signals
My Slack pings. It's Derek Tran from Coastal Therapy Group. He took a demo two weeks ago. His message is a screenshot of a completed onboarding form: he's activated a Clearwater workspace, created his first practitioner account, and connected his patient intake form.
I feel a hit of momentum. Derek represents revenue. At $250 per month, he's already worth fifty dollars of today's ARR. I need four Derek Trans per week to hit my targets.
I Slack back: "Derek, this is huge. Let me grab you at 2 PM and walk through the customization layer so you're not stuck when you try to segment your patient list." He responds immediately with a thumbs up.
I check the Stripe dashboard. Derek's card went through. June's run rate is now $2,850. If the close rate stays at 3-4 per week, I'm tracking toward roughly $150k ARR by September. Not enough. But it's movement.
12:47 PM - Lunch and the metrics check
I order a sandwich and pull up my spreadsheet.
Week to date: 4 qualified leads (down from 6 last week). 2 demos booked (on track). 1 close (Derek). Revenue this week so far: $250.
Last week: $750 from two closes. The week before: $1,200 from three. The pattern is real - I'm closing 2.5 per week on average, which puts me at $10,800 per month. To hit $210k by mid-year, I need to get to at least $17,500 per month by June 30. That's 70 closes by then, or 17 per week.
I'm at 2.5 per week.
The gap makes my chest tight.
But the email approval work is already shifting things. The personalization is higher. The bounce rate dropped from 18% to 11% last week. The reply rate is up. I know what needs to happen: I need to increase volume by at least 60% and tighten close rates. The AI agent is helping with volume. Derek's quick activation and Slack response suggests the onboarding experience is working.
I eat the sandwich.
2:08 PM - Manual work
Derek signs on at 2 PM. We screen-share for thirty minutes. He has a specific question: "Can I run separate campaigns for PT versus occupational therapy?" The answer is yes, but it lives two layers deep in settings. I show him where it is. He takes a screenshot. I know he'll come back in two weeks with a different question, and it'll be in a different place.
I don't have a customer success person. That's me.
By 2:51 PM, Derek's happy and I've created a doc: "Clearwater Quick Start for Therapy Practices." It's not polished, but it covers the most-asked questions. I'll update it as I answer more of them.
Then Michael from Arbor Therapy reaches out via email. He closed three weeks ago, hasn't logged in since the onboarding call. His credit card is still being charged. He's likely to churn.
I phone him. He answers on the third ring.
"Michael, it's the Clearwater guy. I noticed you haven't been in the platform. Everything okay?"
He sighs. "Honestly, I got busy. I'm running five clinicians now and I'm stretched thin."
That's real. I ask him what would unblock him. He says he doesn't have time to set up segments for his different clinic locations. I offer to do it for him on a shared screen tomorrow morning. He agrees.
I don't know yet if I've saved the $250 per month, but I've moved the needle.
3:45 PM - Pipeline review
I look at the three demos I have scheduled for this week. Tuesday afternoon (tomorrow) is Michael's segmentation help. Thursday morning is another practice manager I've been emailing. Friday is Derek's follow-up for advanced settings.
Two out of three of these demos came from outbound that I personally approved. One came from Michael's inbound response. That's the mix I need to sustain: outbound volume on top, inbound interest flowing in.
I see an email from Linda Huang, a practice manager in Portland I'd demoed to six weeks ago. She never closed, went quiet. Today she writes: "I've been watching how our clinic has grown. We need something like Clearwater now more than we did when we talked. Can we schedule a call?"
I write back with a calendar link. Free demo slot Tuesday at 4:15 PM.
That's inbound interest from a warm lead. That's how the math changes.
5:22 PM - Close-out
I've been working since 8:42 AM, not straight but close. I've approved 32 outbound emails. I've handled one customer escalation by phone. I've closed one sale via email. I've prevented one churn by phone. I've done manual customer success work that I'd prefer to automate but can't yet. I've fixed one product issue (the LinkedIn setting). I've reviewed metrics and updated my pipeline.
I haven't written any code. I haven't designed anything. I've sold, operated, and fixed.
I open my todo list and see the recurring theme: "Review agent output before sending" appears every day. "Follow up with warm leads" appears three times. "Handle churn calls" is new. "Document the product for customers" keeps creeping into the backlog.
The AI is multiplying my output. The emails are better because I review them. The close rate is sustainable because I'm handling escalations. But this is not passive. This is not a robot doing everything.
I close the laptop at 6:15 PM. Tomorrow I need to get that close rate to 4 per week. Today I did the work that makes that possible.