What We Hear When We Talk to Finance Teams

The real work behind getting paid.

We spend a lot of time on calls with finance teams. Not selling — listening. Trying to understand what the day actually looks like for the people managing collections at mid-market companies.

What we've learned doesn't match the narrative that most AR software vendors tell. The industry talks about "automating collections" as if the problem is a lack of automation. It isn't. The problem is that the work itself is fundamentally human — and nobody's built tools that respect that.

Here are some of the things we keep hearing.

"I'm crafting emails at 10 PM"

One of the finance leaders we talk to regularly manages collections for a growing company. She handles AR alongside invoicing, vendor payments, and everything else that falls under "finance team of one."

She told us that the reason her collections work well is that she takes the time to actually see the person on the other side. She doesn't send template blasts. She reads the last conversation. She considers the relationship. She crafts a message that sounds like it came from a human — because it did.

The problem is that this takes time. Time she doesn't have during the day, between meetings, between the other 47 things on her plate. So she does it at night.

When we showed her what an AI agent could do — draft a contextually appropriate response that she could review and approve in 30 seconds instead of writing from scratch in 10 minutes — her reaction wasn't "that's cool technology." It was: "I could get my evenings back."

That's the right way to think about what agents do. They don't replace the person who cares about doing the job right. They give that person capacity.

"We just email the customer: what did you pay?"

A fractional CFO we work with serves multiple mid-market clients. He described the cash application process at some of his smaller clients — the ones without sophisticated billing systems.

A payment comes in. It doesn't match any single open invoice. Maybe the customer combined three invoices into one payment. Maybe the amount is off by a few dollars because of a currency conversion or a line item they decided not to pay. Maybe the bank description is so cryptic you can't even tell who sent it.

The solution, more often than he'd like to admit: email the customer and ask what they paid.

He described clerks who handle cash application as people who do well with clear instructions — A, then B, then C. But when a payment doesn't match cleanly, it requires investigation, judgment, maybe a phone call. And that's where things stall. He's seen payments sit unapplied for two weeks. Two weeks where the CFO can't run accurate cash projections, where the aging report is wrong, where the customer thinks they've paid but your books say they haven't.

Two weeks — not because anyone is incompetent, but because the work required to resolve the mismatch takes time that nobody has.

"80% of the work is just send a W-9"

A Director of Billing at a company that's grown through acquisitions told us something that reframed how we think about collections entirely.

She said that 80% of the inbound work her team handles is routine: send a W-9, answer a FAQ, explain why an invoice is higher than last quarter, walk someone through the sales tax breakdown, resend an invoice copy.

Not disputes. Not escalations. Not complex negotiations. Just... answering the same questions, sending the same documents, over and over, to different people at different companies, all day long.

Her team of six — only one of whom is a dedicated collections specialist — spends most of their time on work that requires zero judgment but 100% of their attention. Because you can't batch it. The requests come in throughout the day, from different channels, and each one needs a specific response with the right attachment and the right context.

She knows exactly what she needs: something that can handle the routine requests automatically (send the W-9, answer the FAQ, explain the sales tax) and only surface the exceptions that require a human decision. But the tools she has — her ERP, her collections platform, her outsourced cash application team — don't work that way. They're either fully manual or fully automated with no intelligence in between.

"Is all my money gonna go to AI machines?"

This one came secondhand. A finance leader mentioned that his CEO had said something to the effect of: "Is all my money gonna go to AI machines moving forward?"

It was said in passing, not directly to him — and the CEO who said it wasn't opposed to AI, just wary of cost without proof. It captures a real tension that exists in every mid-market company right now. Leadership sees AI as both the future and a cost center. They know the team is stretched thin. They know they're not going to hire. But they're also not sure what AI actually does versus what it promises.

This is the reality we're building for. Not the innovation-hungry early adopter who wants to try every new tool. The pragmatic finance leader who needs to make a case — to their CEO, their CFO, their board — that this specific tool will pay for itself. Not in theoretical productivity gains, but in actual cash collected, actual hours recovered, actual invoices that would have fallen through the cracks.

That's why we offer a 90-day performance guarantee. We measure the baseline when you start. We track DSO, CEI, collection rates, and aging trends. If the numbers don't move, you walk away. No questions asked.

The last customer who took that deal saw 30x their investment returned in recovered receivables within 90 days. Not because our agents did anything magical. They just followed up — consistently, contextually, at scale — on invoices that the team knew were overdue but didn't have time to chase.

"They're not gonna hire someone new"

A controller at a growing company told us that management had decided not to hire another person for his team. Instead, they moved someone from another department — someone who'd done the job before, which made it less awkward, but also didn't really solve the capacity problem.

There was no discussion of new tools. No budget for software. Just a quiet expectation that the same team would handle more volume.

This is the most common scenario we encounter. It's not that companies don't recognize the problem. It's that the solutions they know about — hiring, outsourcing, enterprise software — are all expensive, slow, or both. So they do nothing and absorb the pain.

The result is a finance team that works longer hours, responds to customers slower, lets invoices age longer than they should, and occasionally drops something important. Not because they're bad at their jobs — because they're doing the work of three people with the resources of one.

"I don't know anything about programming or coding"

One of the things that struck us most was hearing a finance professional say he was more interested in trying Exante precisely because he has no technical background. He saw the instruction editor — where you describe what you want the agent to do in plain language — and said: "You're literally working with a fresh slate. I don't know how to do any of this."

He meant it as a positive. The tool didn't assume he knew what dynamic variables were. It didn't require him to think in templates and conditional logic. It just asked him to describe how he works.

That reaction tells us we're pointed in the right direction. The people doing collections at mid-market companies aren't engineers. They're not workflow designers. They're operators who are expert at their domain — customer relationships, payment patterns, escalation judgment — and they need tools that leverage that expertise without requiring a new set of skills.

What we're building from what we've heard

Every conversation reinforces the same pattern: the work is human. Collections is relationship management. It requires context, judgment, and empathy. The tools that treat it as a sequence of automated emails miss the point.

The teams are small. One to six people handling everything from invoicing to collections to cash application to reporting. There's no slack in the system.

The tools are mismatched. Either too basic (send reminders, generate aging reports) or too complex (enterprise implementations that cost $50K+ and take months to deploy).

The ask is simple. Handle the routine stuff so I can focus on the work that actually needs me. Don't make me learn a new system. Don't make me configure workflows. Just understand how I work and help me do more of it.

That's what Exante is. An AI teammate that handles collections the way you would if you had the time — and asks for your help only when it matters.