4 contractors couldn't keep up. One AI teammate did.
ServiceUp issued 30,000 invoices in 3 months, cut month-end close from 15 days to 2–3, and moved 80 hours a week of contractor work to real accounting. Caitlyn, Controller at ServiceUp, tells the story.
2–3 days
Month-end close (from 15 days)
30,000
Invoices issued in the first 3 months
24 hrs
Invoice issuance (was 2 weeks behind)
80 hrs/wk
Contractor work reallocated
“Once we implemented Ari, I got my life back. That was just a breath of fresh air. I was able to breathe again.”
Caitlyn, Controller at ServiceUp
About ServiceUp
ServiceUp is a car repair management platform. Every closed repair order generates an invoice—customer-facing, high-volume, and coded line by line. AR is the biggest workload in the finance function, and the company's goal was explicit: build a finance team that can take in five times the business without adding five times the headcount.
60-Hour Weeks and a 15-Day Close
Before LedgerUp, ServiceUp's billing scaled the only way manual billing can: with people and hours. Caitlyn started with one contractor covering both AR and AP. As volume grew, the AR workload swallowed the role—40 hours a week, then 50, then 60—and she kept adding headcount until four contractors were carrying the billing workload, because every invoice needed a human to code the customer and line items and push it out.
Every new product and every new customer meant retraining the team. Invoicing ran two weeks behind. Month-end took 15 days to close out invoices, and payment reconciliation took days on its own.
“Over time, as our volume grew, I knew that it just wasn't going to be scalable. The only levers that I could pull at the time was adding contractors, increasing their hours.”
Caitlyn, Controller at ServiceUp
Two Automation Attempts That Still Needed a Human on Every Invoice
ServiceUp didn't jump straight to an AI agent. They tried automation twice first.
The first attempt was an in-house build with the engineering team. When a repair order closed, the system knew which invoice to issue—but a person still had to code the customer and every line item by hand.
The second attempt was an off-the-shelf billing tool built primarily for SaaS companies. ServiceUp's per-repair-order billing didn't fit the mold, so every single invoice still needed a human in the loop to verify the coding and confirm it reached the right customer. The tools changed; the bottleneck didn't.
A Tool You Can Trust vs. One You Have to Babysit
With LedgerUp, ServiceUp brought on Ari—an AI billing agent that follows the team's instructions on coding, issuing, and delivering invoices end-to-end. Caitlyn started cautiously, checking each individual invoice Ari produced. Then she stopped checking, moved to a monthly reconciliation, and found Ari coding correctly and sending to the right customer every time.
“That's just the difference between going from a tool that you can trust to one that you have to babysit.”
Caitlyn, Controller at ServiceUp
Ari knows what it doesn't know
The trust didn't come from Ari being right every time—it came from Ari knowing when it might not be. When Ari hits something ambiguous, it stops, proposes a solution, and asks for approval before taking any action. No silent mistakes to catch at month-end. That behavior is what let Caitlyn go from reviewing every invoice to a monthly spot-check.
“The difference with Ari is Ari knows what it doesn't know, and that has saved us so much time. Ari will stop and ask for an approval and propose a solution before taking any action.”
Caitlyn, Controller at ServiceUp
Everything Changed After LedgerUp Went Live
In the first three months, ServiceUp issued 30,000 invoices through Ari. The KPI Caitlyn cared most about—getting invoices out the door correctly and fast—went from two weeks behind to issued within 24 hours. Month-end close dropped from 15 days to 2–3 business days. Payment reconciliations that took days now take minutes.
The 80 hours a week of contractor time that AR used to consume moved to other pieces of accounting. Caitlyn got at least 10 hours a week back herself—enough to get back to the gym, leave work on time, and stop carrying the close into her evenings.
Before & After
| Metric | Before LedgerUp | After LedgerUp |
|---|---|---|
| Month-end close | 15 days | 2–3 business days |
| Invoice issuance | 2 weeks behind | Within 24 hours |
| Payment reconciliation | Literally days | Minutes |
| AR staffing | 4 contractors at 40–60 hrs/week | 80 hrs/week reallocated to other accounting |
| Invoice review | Human in the loop on every invoice | Monthly reconciliation spot-check |
| Controller's week | Late nights, no gym | 10+ hours/week back |
The Results
- 30,000 invoices issued in the first 3 months — coded, delivered, and reconciled by Ari with a monthly spot-check instead of per-invoice review
- Month-end close cut from 15 days to 2–3 business days — an 80%+ reduction in close time for invoicing
- Invoices out within 24 hours — after running a consistent two weeks behind
- Payment reconciliation in minutes — down from multiple days of manual matching
- 80 hours/week of contractor work reallocated — the AR contractor team now covers other accounting work instead of manual invoice coding
- A finance function built for 5x volume — without adding 5x headcount, which was the goal from the start
Caitlyn's Advice for Finance Teams With Complex Billing
The instinct for most controllers is that their billing is too complex to automate—too many edge cases, too much judgment. Caitlyn's take, after two failed automation attempts and one that worked, is the opposite.
“If you think that your billing process is too complex, you are the perfect candidate to implement this type of process and use AI. You're not implementing the software. You are onboarding a new hire.”
Caitlyn, Controller at ServiceUp
Full Video Transcript
Read the full transcript — Caitlyn, Controller at ServiceUp (3:57)
Once we implemented Ari, I got my life back. That was just a breath of fresh air. I was able to breathe again.
I'm the controller here at ServiceUp. For us, AR is the highest volume. It's customer-facing. We're focused on building a finance function that can take in five times the business without adding five times the headcount.
Over time, as our volume grew, I knew that it just wasn't going to be scalable. The only levers that I could pull at the time was adding contractors, increasing their hours. We had to retrain every time we added a new product or we had a new customer onboard.
I started with one contractor who was focused on both AR and AP, and I gradually had to add a headcount. The contractor that was working on AR kept increasing hours—40 hours, 50 hours, 60 hours a week. But I said, okay, this is too much for her. I need to add additional headcounts. We added two more contractors to focus on AR with her because the process was so manual.
Before LedgerUp, we tried two different versions of automation. The first one had our engineering team involved. So when a repair order closed, we knew which invoice to issue, but we still had to code the customer, the different line items.
The second version of automation, we were using a billing tool. This tool was primarily focused on SaaS companies, so we still needed a human in the loop for every invoice that was getting issued to ensure that it was coded properly and pushed out to the correct customer.
But now, with LedgerUp, we have an AI agent that is following the instructions to a T. So that's just the difference between going from a tool that you can trust to one that you have to babysit.
Initially, I was checking each individual invoice, and then it came to a point where I just kind of stopped checking, and I just did a monthly reconciliation and saw that Ari was coding things properly, sending it to the correct customer. And so, naturally, that just gave me the confidence that Ari was ready to run free.
The difference with Ari is Ari knows what it doesn't know, and that has saved us so much time. Ari will stop and ask for an approval and propose a solution before taking any action.
Everything changed after LedgerUp went live. Month-end, we were taking 15 days to close out invoices. We're doing that within two to three business days now. Payment reconciliations took literally days. We're doing it in minutes.
And I want to say within the first three months, we issued 30,000 invoices, which is kind of unbelievable. And the biggest KPI that I was focused on is, how do we get these invoices out the door correctly and a lot faster? We were two weeks behind, and now we're issuing within 24 hours.
So, with LedgerUp, we've been able to allocate at least 80 hours a week of contractor work to other pieces of accounting. I got probably at least 10 hours a week back.
I found myself dropping the gym. That was something I love to do. So now I get to go back and do that very consistently. I get to go home a little bit earlier and spend time with my husband, and giving myself time to just relax and reboot.
If you think that your billing process is too complex, you are the perfect candidate to implement this type of process and use AI. You're not implementing the software. You are onboarding a new hire.
Company
ServiceUp
Industry
Car Repair Management
AR model
High-volume, per-repair-order invoicing
Team
Controller + AR contractors