8:42 AM - Inbox triage
I open Pylon on my laptop in the Reyes Family Practice parking lot. I'm still sitting in the truck, coffee in my hand, waiting for the office to open. My phone lit up twice already this morning: a Slack alert at 6:15 when Marcus Rodriguez's trailer lost GPS signal somewhere outside Albuquerque (turned out to be bad cell coverage, not a breakdown), and a Gmail notification from Tom Burgess at Trinity Logistics asking me to call him back about a new integration.
The dashboard is clean. Twelve loads assigned yesterday are either in-transit or delivered. Zero red flags. That's the moment that sells you on Pylon: you open it and you see your whole operation on one screen. No hunting through email chains. No calling drivers to ask where they are.
Three new signups in the intake queue: two from Facebook group leads I posted to last Thursday, one from a referral. I add a note to schedule onboarding calls for Thursday. The agent has already flagged that one of them - Linda Paulson at Paulson Trucking - runs a lot of hazmat loads. I make a note to walk her through the CDL endorsement matching rules that day.
I check Slack. No new alerts in the past hour. Good. Revenue dashboard shows 2.1 loads per driver assigned yesterday, up from 1.8 two weeks ago. The math feels real now: each tenth of a load is someone's fuel money, someone's dispatch time. Last Tuesday I hit 1.9 and it felt like failure. Today 2.1 feels like progress.
10:15 AM - Reviewing agent output
Back in the office, I pull up the email drafts. The agent prepared a response to Carol Reyes at Reyes Family Practice. Carol is one of my anchor customers - been with me since month one - and she'd emailed yesterday asking about our new SLA language. She's a cautious customer. She reads contracts. She calls me when she sees something that doesn't align with how she thinks about the business.
The agent's draft is technically accurate. It covers the three points she asked about: what happens if we go down for maintenance, how long dispatch takes once a load hits the system, what escalation path she has if something breaks. But the tone feels corporate. Distant. Carol runs a family practice, not a call center. She knows I'm one person. She expects me to know her preference for afternoon loads over morning loads because I pay attention.
I rewrite it. I keep the agent's substance but swap the cadence. I add a sentence about the new walk-in clinic expansion she mentioned last month, and I tell her I'm blocking Tuesdays and Thursdays for her now because that's when her staff hours are tightest. I add it to my Outlook. This is the work I can't delegate yet: knowing my customers as humans, not just load assignments.
I send it manually. The agent learns from what I write, but I need to be the one who hits send on the stuff that matters for keeping Carol.
11:45 AM - Edge case and a flag
While I'm still on email, a notification drops in Slack. The agent has flagged an exception in the dispatch logic. Marcus Rodriguez has 3 hours of logged drive time remaining today, and a 480-mile load just came in with a 4-hour window. The agent is asking me to make the call: do we assign it to Marcus under 34-hour restart rules and risk putting him over if traffic hits, or do we reject the load and bounce it back to the broker.
This is the thing nobody talks about when they sell you on automation. The hard decisions don't disappear. They just get faster. I pull up Marcus's logs in the ELD feed. He's been steady all week - nothing reckless. I check the pickup location: 40 miles from where he is now. The window is 2 to 6 PM. I look at traffic. Usually clear on 40 at this hour.
I do what I've learned to do: I slack Marcus directly and ask him. Not the agent asking. Me. I write: "Marcus, 480-mile load Albuquerque to Odessa, pickup between 2 and 6. You'll land right at 11 hours before restart. Weather looks clean, but 40 can get hairy. What's your read."
He comes back in four minutes: "I'll take it if you think the load quality is good."
So I check the broker. It's Carol Reyes. Known rate. Known receiver. I authorize the agent to dispatch it and SMS Marcus the confirmation. The load gets assigned. Marcus gets moving. This is what good feels like: a decision made because I know my driver and I know my customer and I have the data in front of me. Not a robot playing probabilities.
12:30 PM - Lunch and the metrics check
I grab a sandwich from the deli down the street and open the metrics dashboard. Week-to-date: 18 loads assigned, 14 delivered, 1 still in-transit, 1 pending pickup, 2 rejected by drivers (both Marcus and another driver, both valid reasons). The 77.8% delivery rate is better than my personal average when I was still driving. That's not something I calculated; that's just what the board tells me now.
Revenue: 1,847 dollars this week so far. If I keep this pace, I hit about 7,400 by Friday. My target for June was 6,500. I'm going to beat it by almost 15 percent. That still doesn't sound like much until you remember that my first month in operation was 3,200 total.
I check the Stripe dashboard. One failed payment - a payment from a new dispatch partner I onboarded last week. Their credit card got declined. I don't know if it's an issue on their end or if they forgot to update it. That's a manual thing. I make a note to call them. This is the work I didn't expect: somebody's gonna forget their payment method, and I'll have to ask them to update it.
I also notice a churn notification from the system. A customer from my first cohort, DeShawn Fuller at Fuller Motor Freight, hasn't logged in in 11 days. The agent has flagged this. His contract ends June 15th, and if he's not logging in, he's probably not thinking about renewal.
2:08 PM - Escalation and a hard conversation
I decide to call DeShawn before the end of day. I know what's probably happened: he tried to assign a load to a driver who wasn't certified for it, got stuck, got frustrated, and decided Pylon was too complicated. It's happened twice before. Both times it was a misunderstanding about hazmat rules. Both times I walked them through it and they came back.
DeShawn picks up. I ask him straight: "You've been quiet for almost two weeks. Everything alright with the system."
There's a pause. Then: "Honestly, I'm not sure the tool is saving me time yet. I'm still doing a lot of manual work on the back end."
He's right. And he's early-stage. Most of my customers don't feel the savings until month two or three, when they stop second-guessing the agent and trust it more. But I can't tell him to wait it out. I can only show him the work I've cut out of his day. I walk him through his own numbers: in March he logged in an average of 2.3 times per day; last week he averaged 1.6. He's spending less time in the system. He just doesn't feel it yet because he's still doing the email thing.
We talk for 15 minutes. I show him the draft email the agent prepared for Carol Reyes and the conflict resolution I'd done with Marcus. I explain that this is the moment most people give up, but it's the moment where the system starts learning his customers and his drivers. He asks good questions. I ask him when his next big onboarding week is and tell him I'll call him Wednesday and do an hour-long walkthrough of how to handle it with Pylon.
He doesn't cancel. He says he'll stick it out. I tell him I'll check in Thursday. I know this is the toughest part of my job right now: retention doesn't happen through the tool alone. It happens when the owner picks up the phone.
4:30 PM - Pipeline and systems check
I'm back at the office. I pull up the Linear board where I track feature requests and bugs from customers. One request from Maria Ochoa at Mountain Valley Dispatch about better visibility into driver HOS time remaining. One bug I reported to myself last week about the SMS delivery confirmation sometimes coming in late.
I fix the bug. It's small: the agent was batching the ELD pull once per hour instead of every 15 minutes. I adjust the polling frequency in the Settings tab. Takes five minutes. I write a manual note back to Maria explaining that we're working on HOS visibility but in the meantime she can pull each driver's detail card. I offer to show her on a call tomorrow.
I check the demo calendar. I have three onboarding calls scheduled for Thursday and two demos set for Wednesday afternoon. One of the demos is with a small parcel dispatch outfit in Denver. They're bigger than my ICP - probably 25 drivers - but they came through a referral and their current system has been a mess for years.
6:15 PM - Wrap
I close the laptop around 6:15. I review the day in my head as I'm packing up.
Three new leads. One customer at serious churn risk who I talked off the ledge. One edge case that took good judgment and a conversation with Marcus. One small bug I fixed myself. One email I rewrote because the tone mattered. One failed payment I have to chase down. One customer metrics review that showed real progress. One new integration conversation I still need to schedule.
This is not passive income. This is not the agent doing the work. This is me using the agent to see the whole picture faster, to catch the problems before they spiral, to know which decisions need my judgment and which ones the system can handle.
Last month, a dispatch day felt like chaos. Today it felt like work. But it's work I built, and it's work that's scaling. If I can keep the retention rate up and push my load volume to 3 per driver per week, I'll be looking at 11, 12 thousand a month by August. That's the number where I'd actually feel like I could hire a part-time dispatcher and step back.
Right now I'm still the dispatcher. Pylon is just making me a much better one.