For the latest news on Financial Operations Automation

The out-of-office replies are piling up, but late payments don't take a summer break. This month we're looking at what still works once the team's out of the room and the agent is running the show:

• We examine why a single identity check is no longer enough to stop fraud, and what a fraud function needs to fall back on once fraudsters know how to subvert ID verification.
• We sit down with Upflow's CEO Alex Louisy to unpack the one thing that separates a collections agent you can demo from one you can actually hand your book to.
• We include our usual round up the most interesting things we've been reading in AR - perfect beach material!

Hackett Group provides first quantification of benefits of end-to-end AI automation of O2C cycle

End-to-end AI automation can deliver:

52–59% reduction in Order-to-Cash costs

83% more time talking to customers instead of chasing data

85% reduction in days delinquent

Source: The Hackett Group

A report from the Hackett Group last week provided the first attempt we've seen of a comprehensive quantification of the benefits of automating the entire order-to-cash cycle. One can quibble about the methodology and the precise numbers. But the evidence that AI automation can both cut costs and improve outcomes in AR is compelling.

IDV alone isn't enough any more. And that's not a reason to stop using it

Three prospects or clients came to us this month with the same story: identity fraud that walked straight through their verification flow - despite ID Verification (IDV). Some cases were sophisticated real-time injection attacks, others just a forged document that shouldn't have made it past a basic check. Either way, it's not bad luck, it's a trend. We break down why relying on a single defence is now inadequate and what a broader, always-on fraud function looks like instead.

Read more

Building collections agents that work: an interview with Upflow's CEO Alex Louisy

"If you can't see the invoice, the payment status and the contact history, you don't have an agent. You have a demo."

— Alex Louisy, Co-founder & CEO, Upflow

Barely a week goes by without another vendor announcing an "autonomous collections agent." Alex Louisy, co-founder of Upflow, has spent eight years building AR automation and has heard every version of the pitch. We sat down with him to ask the question most of the marketing skips: what actually separates an agent you can trust with your book from one you can only trust with a demo. His answer starts with a hierarchy of needs in AR, and ends somewhere most vendors haven't got to yet.

Read more

What we've been reading this month

1. We enjoyed this Notion survey of AI use across 6,000 companies. While 49% of company leaders say they are confident in their firm's ability to deploy AI, this view is shared by only 23% of their employees. It's a helpful reminder that there's a very big difference between "we had a great demo in Exco last week" and "our staff are actually more productive now" (here)

2. We've been talking for a while about how the first use cases of AI agents talking to each other is likely to be AR and AP. So it was great to see Maex Ament's new venture, Causa Prima, which is built around precisely this idea. As one of the founders of Taulia (now SAP's working capital solution) Maex is a real visionary, so it's always worth following what he's up to.

3. On a lighter note, we enjoyed this piece of AI hype about an AI agent startup whose agent allegedly raised a $100m equity round completely on its own. Not sure if this says more about how capable the agent was, or how desperate VCs are to throw money at "anything AI" these days!

If you'd like to receive this newsletter direct to your inbox, sign up here