New report: The 2026 State of AI in Finance

Auditoria logo in multi-color

Finance Can’t Buy AI on Faith Anymore: How to Measure AI ROI in AP and AR

Carey Hite November 3, 2026

AI value isn’t automatic. Learn three things finance leaders need to know about proving AI value in AP and AR, from baselining your process to choosing the right partner.

Key takeaways

  • AI value in finance is not automatic. Apollo’s chief economist wrote on October 24, 2026 that AI has “delivered zero financial payoff to corporate America outside of tech,” so finance leaders should expect to prove value, not assume it.
  • Measure before you buy. Baseline cost per invoice, invoice cycle time, touchless rate, inquiry volume and response time, and days sales outstanding (DSO) before you talk to a vendor.
  • Set targets the business cares about. “Save time” is not a target. “Cut invoice cycle time from nine days to four” is.
  • Choose a partner who will be measured with you: a team that owns outcomes after go-live, reports against your targets on a regular cadence, and sets realistic expectations about ramp time.

Recently, Apollo’s chief economist, Torsten Slok, made a blunt claim: AI has “delivered zero financial payoff to corporate America outside of tech.” (Fortune) You can argue with his metric, but finance leaders shouldn’t ignore the message: AI value isn’t automatic, and “we bought it” is no longer an answer to “what did we get?”

For the past two years, a lot of AI got bought on momentum and FOMO. The board asked about it, a competitor made some headlines, a vendor demo looked impressive. That window of pilots and guessing is closing fast. CFOs are now being asked a much simpler question: what did we actually get for it?

I lead Professional Services at Auditoria, so I live on the other side of that question every day. My team implements AI in the Office of the CFO across their AP and AR functions, and we are accountable for whether it moves the numbers. Working with enterprise finance teams every day, I want every leader to understand these three things.

1.  Why isn’t “everyone else is doing it” a business case for AI in finance?

Buying software because it has AI in the name, or because leadership said to, is no longer a defensible decision. Budgets are tighter, and AI spend now gets the same scrutiny as every other line item.

That’s a good thing. AI in finance can deliver real value: fewer manual touches, faster cycle times, a better experience for vendors and customers. But value you don’t measure can’t be defended at renewal, can’t justify expansion, and won’t earn the trust of the team that has to use the tool every day.

The organizations that start tracking value now will still be investing in AI two years from now. The ones that don’t will be explaining why the pilot stalled.

2. How do you measure AI ROI in finance? Start inside your own four walls

I’m not saying you have to go build a comprehensive digital representation of your entire organization before you can realize value, but before you talk to a single vendor, figure out where value is hiding in your own operation. That takes an honest internal assessment:

  • Map the work. Which processes eat the most hours? In AP and AR, it’s usually invoice entry and coding, exception handling, vendor and customer inquiries, and collections follow-up.
  • Baseline it. Capture today’s numbers: cost per invoice, cycle time, touchless rate, inquiry volume and response time, DSO. If you can’t measure the before, you’ll never prove the after.
  • Set targets the business cares about. “Save time” isn’t a target. “Cut invoice cycle time from nine days to four” is.
  • Check your foundation. AI runs on your data and inside your ERP. Messy vendor data or undocumented approval rules will cap your results, no matter which tool you pick.

This step is unglamorous, and it’s the one most often skipped. It just so happens to also be the most important. In my experience, this separates successful programs from those that fail.

3. How do you evaluate an AI vendor for finance? Choose a partner who will be measured with you

Once you know where to aim, the vendor conversation changes. You stop asking “what can your AI do?” and start asking “how will you help me prove it worked?” A good partner will:

  • Bring a team whose job is value, not just go-live. Look for a dedicated counterpart to whoever tracks results on your side, someone who owns outcomes after the implementation team moves on.
  • Report against those targets on a regular cadence, and tell you when something is off track and what they’ll do about it.
  • Set realistic expectations. AI gets better as it learns your vendors, your rules, and your exceptions. Expect a ramp, not a light switch; the most meaningful metrics often take several months after go-live to show up.

At Auditoria, that’s the job of our AI Outcomes team. We engage during implementation and stay with customers well past go-live, tracking value against the goals we set together.

If a vendor can’t tell you how they’ll measure success, they’re asking you to buy on faith.

Get the toolkit: join the conversation on measuring AI value

This is the short version. The days of implementing AI for the sake of AI are ending, and the finance teams that win will be the ones who can prove what it’s worth.

On November 10 at 11 am PT / 2 PM ET join our webinar, Desperately Seeking AI Success Metrics, for the deep dive: the process components to assess, how to evaluate AI vendors, and a practical toolkit for realizing value from AI in finance.

Frequently asked questions

1. Does AI deliver ROI in finance?
It can, but not automatically. Value shows up as fewer manual touches, faster cycle times and a better experience for vendors and customers. Teams that baseline their numbers and track them against targets can defend the spend at renewal. Teams that don’t have no proof either way.

2. What metrics should finance teams track to measure AI value in AP and AR?
Track cost per invoice, invoice cycle time, touchless rate, exception rate, inquiry volume and response time, and days sales outstanding (DSO). Capture each one before go-live so you can prove the change afterward.

3. How do you measure the ROI of AP automation?
Baseline your AP metrics before implementation, set a target for each one, and compare actual results against those targets on a regular cadence after go-live. Allow for a ramp, because the most meaningful metrics often take several months to show up.

4. What should I ask an AI vendor before buying AP or AR automation software?
Ask who owns outcomes after go-live, how and how often they report against your targets, what they do when a metric is off track, which metrics take months to show, and what they need from your data and ERP.

5. How long does it take to see results from AI in finance?
Expect a ramp, not a light switch. AI gets better as it learns your vendors, your rules and your exceptions, and the most meaningful metrics often take several months after go-live.

6. How can finance teams reduce DSO with AI?
Baseline DSO and the hours your team spends on collections follow up and customer inquiries, then set a DSO target. AI agents that run AR work from the inbox to the ERP within your policies can take on follow up and inquiries so your team focuses on priority accounts.