Billtrust CTO says finance AI gains depend on cleaner workflows

Finance teams are unlikely to capture the full value of artificial intelligence if their data remains split across enterprise systems, spreadsheets and vendor tools, Billtrust Chief Technology Officer John Landy told PYMNTS. The near-term effect for chief financial officers is operational rather than speculative: AI can accelerate finance work, but it also exposes weaknesses in the architecture beneath it.

Landy said companies risk limiting returns when they attach AI to fragmented systems after finance processes have already been built. In his view, the work starts with data design, control and context across the finance stack, rather than with model selection alone.

“The biggest issue is around having something be a bolt-on investment after the fact when you have fragmented data and solutions and systems and vendors working in your current environment today,” Landy told PYMNTS. He said companies that rethink architecture “from the data all the way up” are better positioned to benefit from AI investment.

Data quality sets the ceiling

Landy said AI tools are now widely accessible inside many organizations, which makes the quality of the underlying data more decisive. “It will present the data however accurate you give it,” he told PYMNTS.

Accounts receivable illustrates the problem. According to Landy, many finance departments rely on several enterprise resource planning systems, payment platforms, customer records, outside vendor applications and spreadsheets. Some data flows in real time, while other information arrives through batch updates or manual reconciliation.

That mix can create timing gaps. Landy described spreadsheets as point-in-time records that staff then try to match against live systems, batch systems and reconciliation workflows. Adding AI above those processes may speed analysis, but it does not by itself make the data consistent.

For accounts receivable, the mechanism is straightforward. If invoice-to-cash data is structured and available where finance staff already ask questions, AI assistants can answer plain-language queries without requiring employees to open separate reports or move data into another application. Landy’s argument is that the value comes from placing reliable context into the workflow, rather than treating AI as a separate destination.

Finance shifts from reporting to earlier intervention

Landy told PYMNTS that the most useful productivity gain may come from reducing the time employees spend collecting information before making decisions. “If you can save the time people spend assembling information, that is the biggest difference,” he said.

That changes how finance teams respond to events. Landy gave the example of a regular customer payment that does not arrive when expected. A connected system could identify the break from normal payment behavior, assess the effect on cash forecasting and alert the relevant employee before the delay becomes a collections issue.

“If you can detect in real time that a payment has not arrived when it typically does, you can react and have a more meaningful conversation before it becomes a problem,” Landy said.

Automation still needs limits

Landy said some finance workflows are better suited to high levels of automation than others. Cash application, invoice matching, aging analysis and fraud detection are candidates because they involve large volumes, repeatable patterns and measurable results. In those cases, employees may focus mainly on exceptions.

He drew a line around decisions involving important customer relationships, large disputes, legal exposure or reputational risk. “If you have an important relationship, you are not going to want the process automated to the point where no one is available to handle the communication,” he told PYMNTS.

Landy framed the governance issue as a choice among workflows where humans are directly involved, workflows where humans supervise, and workflows that can run with more autonomy. He said decisions involving legal, brand or reputational risk should keep a person in the process.

The shift also changes the relationship between finance and technology teams. Landy said CFOs do not need to become machine-learning specialists, but they do need to understand how financial data is stored, secured, shared and connected to AI tools.

“The number one requirement is a partnership between CFOs and their IT teams to ensure the right infrastructure is being evaluated and addressed,” Landy told PYMNTS.

This story draws on original reporting from PYMNTS.