By Alex Mifsud

Like every other business function, Accounts Payable is having its AI moment – and understandably so.

Everywhere you look, finance and operations teams are experimenting with automation, predictive analytics, and machine learning. Recent surveys show that 57% of accounting and bookkeeping firms plan to invest in AI in the next year, and almost 98% have already used it for tasks like data entry, reconciliation, and fraud detection1,2. This shift is no longer confined to innovation labs or future-of-finance panels. It’s well underway.

With all this intelligence, can Accounts Payable become a fully automated process where invoices simply get paid, just at the right time to optimise on cash management, discounts, supplier relationships and so on.  If not, why not – where does it still fall short?

That’s the question I find myself circling. Because for all the power of AI to parse data and work out optimal strategies against goals, most AP platforms still leave the final step – actually executing payments – to an entirely separate system. It’s as if your satnav finds the best route, but then asks you to find another app to give you directions while driving.

So what happens when insight and action live in the same place?

The AI Wave in AP – and What It Hasn’t Solved

AI is already solving pain points that have dogged AP teams for years. It auto-classifies invoices, checks them against purchase orders, flags anomalies, and can even forecast cash flow. We’re now seeing intelligent categorisation, smart matching, and risk scoring rolled out in mainstream software.

But the moment it’s time to pay – to initiate a bulk supplier run, the core task of many AP teams – users are still switching context. From AP systems to spreadsheets. From approval platforms to bank portals. From “this looks right” to “I hope I don’t mess up this BACS file.”

Why do we tolerate this gap?

Embedded Payment Run as the Missing Piece

Some AP platforms are beginning to close the loop. Paperchase, a hospitality-focused accounts payable platform, recently introduced Embedded Payment Run via Weavr, enabling finance teams to handle bulk supplier payments directly within their workflow – no file downloads, no switching to bank portals. It’s a notable first: the ability to go from invoice approval to multi-supplier payment execution, all in-platform3.

Others are moving in a similar direction. Tipalti, for instance, offers embedded payments as part of a broader global payables suite4. And platforms like Finway are integrating card-based spend management within their AP and expense tools5. But Paperchase’s move into fully embedded bulk payment execution stands out as a signal of where the AP category may be heading.

What’s significant about these examples isn’t just that payments are now possible inside the platform. It’s that execution no longer breaks the flow. And when paired with AI, it opens up entirely new ways of working.

What Could AI + Embedded Payment Run Actually Look Like?

Let’s step into a couple of practical scenarios:

✅ Scenario 1: AI-Initiated Bulk Payment Batches

Today, finance managers often spend hours compiling pay runs: checking due dates, validating supplier bank details, grouping payments, and exporting batches to the bank. It’s tedious. It’s high stakes. And it happens every week.

Now imagine this instead:

  • AI reads approved invoices and flags any that are invalid, duplicate, or outside normal thresholds.
  • The system groups valid payments based on due dates and cash flow rules.
  • A draft bulk payment is auto-populated – and the finance lead just reviews and clicks approve.
  • Payment is executed inside the AP platform, without logging into a bank.

Same controls. Less overhead. Zero switching.

🚨 Scenario 2: Payment Block for Suspicious Vendor Activity

Now imagine you’re processing a bulk run to dozens of suppliers – but AI flags a new vendor with slightly mismatched account details.

Instead of the whole batch stalling or a risky payment slipping through, the system does two things:

  • Automatically pauses the specific transaction and notifies the team for review.
  • Allows the rest of the pay run to proceed uninterrupted.

The payment execution lives inside the platform. Which means the response – the remediation – does too.

And That’s Just for Starters

Once the tools are in place for seamless execution, the next step is for AI agents to do the bulk of the work.   It’s early days still, but the SaaS model is being disrupted by Agentic AI, where the human actions supported by SaaS, in terms of presenting data or steps in a workflow, are taken over by AI agents that interpret the data in the context of the business, and perform actions accordingly.

Agents, however, need the right tools to do their thing.  Right now, payments systems – and the regulations that govern them – have not been designed for use by agents and the industry, including regulators, has yet to catch up.  When regulations allow, Embedded Payment Run, will enable agents to do the whole job. In the meantime, Embedded Payment Run enables agents to nudge the authorised human for a simple single approval – in line with strong customer authentication regulations – before going on to execute the payments involved in an Accounts Payable session. 

Closing the Loop

It’s often said that AI will be the brain of tomorrow’s finance function. I’d argue that execution is the hands. And just like in the human body, it’s the connection between the two that allows for real action.

As AI continues to mature, the question won’t be whether platforms can advise us wisely – it will be whether they can carry out those decisions seamlessly, safely, and in context.

That’s the quiet convergence we’re starting to see in AP: systems that don’t just inform, but complete. And I suspect we’ve only scratched the surface of what that will make possible.

Sources

  1. “Accountant Technology Survey 2024 Report.” Firm of the Future.
    https://www.firmofthefuture.com/news/accountant-tech-survey-2024/
  2. “Accountant Technology Survey 2024 Report.” QuickBooks Resource Center.
    https://quickbooks.intuit.com/r/small-business-data/accountant-tech-survey-2024/
  3. “Weavr Launches Embedded Payment Run with Paperchase.” Weavr Blog.
    https://www.weavr.io/blog/weavr-launches-industry-first-embedded-payment-run-technology-debuting-with-paperchase/
  4. “AP Automation: End-to-End Accounts Payable Software.” Tipalti.
    https://tipalti.com/products/ap-automation/
  5. “Weavr Partners with Finway to Launch Embedded Cards for SMEs.” Weavr Blog.
    https://www.weavr.io/blog/embedded-finance-provider-weavr-partners-with-finway-to-issue-physical-and-virtual-cards-for-smes/