A supplier promises to deliver parts on the 18th of the month. The shipment gets pushed to the 25th. The enterprise resource planning (ERP) system updates the expected delivery date, and when the goods finally arrive, the supplier is recorded as having delivered on time.
The dashboard is green. The factory waited an extra week.
“Finance will see something, procurement will see something, and production will see something else,” Axya Founder and CEO Félix Bélisle-Dockrill said in a discussion hosted by PYMNTS CEO Karen Webster.
For chief financial officers, the consequences extend beyond purchasing efficiency. Inaccurate delivery records can distort supplier evaluations, obscure contractual obligations, complicate invoice reconciliation and weaken the information used to make working capital decisions.
“Releasing payment if something is perceived as on time, but it’s seven days late, there may be contractual consequences,” Webster said, connecting the issue directly to payment controls.
As a result, large enterprises may be managing supplier performance using data that systematically obscures the very disruptions they need to prevent.
The larger opportunity is not just automating procurement but creating a more reliable record of what suppliers actually commit to delivering, Bélisle-Dockrill said. Axya raised $17 million Canadian dollars (about $12.1 million) in September to do just that.
The Purchase Order Becomes a Source of Financial Intelligence
Enterprise procurement remains dependent on emails, PDFs, spreadsheets and disconnected software systems. Human employees typically manually bridge these information silos.
“The buyers are the human routers between all these different stakeholders,” Bélisle-Dockrill said, adding that these “human routers” frequently manually transfer supplier confirmations and changes into internal systems, and those changes do not always reach finance.
The problem becomes particularly consequential in complex manufacturing, where a single purchase order may contain delivery commitments, pricing agreements and specifications that change repeatedly before fulfillment.
“A buyer will send a purchase order with expected delivery date and prices,” Bélisle-Dockrill said. “Supplier will confirm something else, but because it’s lost in emails, it might not be captured. When finance is ready to pay the invoice, it doesn’t match what they were expecting.”
The next frontier of procurement artificial intelligence may not be purchasing goods faster. It may be ensuring that the information companies use to price, finance and manage those purchases reflects what is actually happening across their supply chains.
AI Gives CFOs Back the Procurement Data Supplier Failures Hide
We’d love to be your preferred source for news.
Please add us to your preferred sources list so our news, data and interviews show up in your feed. Thanks!
If the financial opportunity is significant, adoption remains difficult. Large manufacturers operate fragmented ERP environments, often combining legacy on-premises infrastructure with newer applications and supplier-specific communication channels.
“Inertia is built into the business process itself because switching comes with its own cost,” Webster said.
Bélisle-Dockrill estimated that tactical activities consume 90% to 95% of procurement professionals’ time, leaving limited capacity for supplier development, contract optimization and risk management. Recognizing that reality, Axya’s owns approach minimizes disruption by integrating with existing inboxes and ERP systems while allowing suppliers to communicate through familiar channels.
“The main pieces is no change, a little bit of value, a little bit of change, more value,” Bélisle-Dockrill said, adding that enterprise AI adoption may depend less on technological sophistication than on how seamlessly it fits existing operations.
Procurement AI’s Next Frontier Is Decision-Making
The workforce implications extend beyond efficiency. As experienced procurement professionals retire, companies risk losing institutional knowledge that rarely exists in formal systems. AI could help preserve that expertise while allowing less-experienced employees to assume greater responsibility.
Recalling the departure of a senior colleague, Bélisle-Dockrill said, “My productivity dropped by five to tenfold.”
As a result, the next competitive advantage of deploying AI in procurement may come from moving beyond identifying supplier problems to determining which actions should follow from them.
Bélisle-Dockrill said his firm is developing capabilities that distinguish decisions suitable for AI agents from those requiring human oversight, with broader deployment targeted for 2027.
For CFOs, identifying a delayed shipment improves visibility. Determining whether to renegotiate delivery terms, adjust payment expectations or escalate supplier risk begins influencing working capital.
Watch the full PYMNTS interview with Axya CEO Félix Bélisle-Dockrill and PYMNTS CEO Karen Webster to hear more about:
- Why procurement data may be hiding costly supplier failures. Bélisle-Dockrill explained how disconnected systems can mask late deliveries and price changes, leaving CFOs with misleading performance metrics, payment disputes and unexpected costs.
- Why the best enterprise AI may be the kind nobody notices. Webster and Bélisle-Dockrill discussed how embedding AI into existing inboxes and ERP workflows can overcome adoption barriers while freeing procurement teams for higher-value work.
- Why procurement AI’s biggest payoff may come before the invoice. Bélisle-Dockrill said he sees AI moving from tracking supplier performance to recommending actions, helping finance teams anticipate disruptions, protect working capital and make better purchasing decisions.
For all PYMNTS B2B and AI coverage, subscribe to the daily B2B and AI newsletters.