<style> .responsive-heading { font-weight: bold; font-size: clamp(1.2rem, 2vw, 2rem); line-height: 1.2; margin: 1em 0; } </style> <div class="responsive-heading"> Please Try a Different Browser </div> <p>You are using an outdated browser that is not compatible with our website content. For an optimal viewing experience, please upgrade to Microsoft Edge or view our site on a different browser.</p> <p><strong>If you choose to continue using this browser, content and functionality will be limited.</strong></p>

PYMNTS What’s Next In Payments with Emily Nash-Walker

The Last Human Touchpoint in the Payment Loop: Which Approvals Will Survive AI?

Most of the invoice-to-pay process already runs without anyone touching it. Invoices are captured, matched and validated. Exceptions get flagged. Batches build themselves. But at the end of the line there’s usually still a person with a finger on the button: releasing the payment over a certain threshold, paying a first-time supplier, clearing the exception nobody has written a rule for. Sometimes the reason is regulation or liability. Sometimes it’s the dollar amount. And sometimes it’s simply that nobody has been willing to take the human out yet.

In this PYMNTS TV interview, Emily Nash-Walker clarifies that the question isn’t whether humans stay in the loop. It’s which loop, why that one and for how long. The opportunity for AP and finance leaders is to separate the work of preparing a payment from the authority to release it, and then let confidence and risk decide where people still add value.

It's important to look at where you need control and how you can automate that. Emily Nash-Walker, Head of AP and AR, Tungsten Automation

Watch the full interview to learn:

  • Where control really exists. Upstream AP can be highly automated while the release of funds stays a separate control point. That control is shaped by transaction value, payment type, supplier history, country exposure and fraud signals./li>
  • Why the approval is still human. Regulation, liability, trust and data quality all play a part. Some of it is just organizational habit.
  • What would have to change. Complete, connected invoice, supplier, PO, receipt and payment data, plus decisions you can explain, are what make finance teams comfortable letting automation act.
  • What replaces the person. It isn’t full autonomy. Work is sorted by confidence and risk: high-confidence payments that follow policy go straight through, and everything else comes up as an exception with the context needed to decide.
  • Who owns the mistake. Automation doesn’t remove accountability. It makes decision rights explicit. Finance and treasury leaders set risk tolerance, approval limits and escalation paths, and audit trails show what was decided and why.

Related resources