eHealthNews.nz: AI & Analytics

Health NZ sees major efficiency gains with AI-powered accounts payable

Wednesday, 16 April 2025  

FEATURE - Industry Innovation Article - Fujifilm

Health New Zealand | Te Whatu Ora has successfully rolled out an AI-powered accounts payable solution nationwide, bringing significant efficiencies by replacing manual data entry, reducing errors, and improving supplier communication.

The Esker platform from FUJIFILM Business Innovation NZ handles data extraction, coding, matching, and PO-based invoice reconciliation. 

The platform integrates with Health NZ’s Financial, Procurement, and Information Management (FPIM) system, which brings all districts and entities under a single national financial platform.

 

A rapid implementation

Mel Patrimonio, Health NZ financial transactions manager, says the rollout of Esker has been rapid. 

Initially implemented in the Northern region – Counties Manukau, Te Toka Tumai Auckland, Te Tai Tokerau Northland and Waitematā - in August last year, it was quickly expanded nationwide. 

“We rolled it out very quickly to all districts and shared services entities, and had one final district go-live in March, so now the whole organisation is using Esker as a tool for accounts payable, which helps to create more consistent AP processes across the country,” she explains.

The transition was particularly significant for districts that previously had no scanning solution at all. 

“Shifting from paper-based or manual data entry to an automated system required a change in mindset. Some staff adapted quickly, while others needed more training and support,” she says.

“We are now at a point where most people have adjusted. It took years to implement the previous system in 2017, but this transition has been much faster, and I am really pleased with how quickly it has progressed.”

 

Realising benefits

Health NZ processes 100,000 non-EDI invoices monthly, with a month on month growth in volume. 

Patrimonio says the primary benefit of Esker is its ability to handle increasing invoice volumes without having to necessarily add more resources. 

“We know that volume increases over time, and we cannot continually put more resource into manually entering the invoices into FPIM,” she says. 

“A lot of our invoices have multiple lines—some have hundreds—and trying to match them manually takes a lot of time.
“The other part of that data entry is we have to attach the
invoice image into our financial system. This is all manual, but this solution actually automates that.”

Patrimonio says automating data capture has reduced manual entry errors and allows her team to focus on managing supplier queries and end-user support rather than just entering data.

“Even entering a date manually carries the risk of mistakes. Automating data capture solves that issue as well,” she says. 

The system also ensures invoices are matched correctly with supplier details.

“It checks our master data, ensuring we are paying the right supplier. If there is any variance in bank account or GST number, the accounts payable clerk can check it.”

She adds that Esker has also streamlined communication with suppliers. 

“Before, if there was an issue with an invoice, it was difficult to track emails and responses. Now, we have better communication with suppliers because everything is tracked within the system, as the history is captured in there,” Patrimonio explains.

“It also gives you insight in terms of what is happening with the supplier, because it offers real time reporting and tracking tools.”

This means the team can go back to suppliers directly about any issues to make the process easier for everyone involved.

 

The power of AI

Another significant advantage is the solution’s AI-driven learning capabilities. 

“The AI learns from the corrections we make: you teach it by using your cursor to say ‘this is actually where these fields are’, so it is recognising invoice fields - such as purchase orders, GST numbers, and bank account details. Our goal is to have invoices auto-post without manual intervention,” she says. 

However, careful training is required to ensure the AI learns correctly. 

“If someone teaches it the wrong thing, it will learn the wrong thing, so we focus on proper training.”

As the AI continues to learn and improve, Health NZ expects further efficiency gains. 

“Technology is here to make things easier, and this system has helped solve and highlighted the issues we had before. It is all about making processes smoother and ensuring invoices are processed quickly and accurately,” Patrimonio says.

 

Future plans

Irene Direnko-Smith, head of process automation, FUJIFILM Business Innovation NZ, says its subscription model, based on volume, allows for scalability and cost management, meeting Health NZ's requirements. 

There are also possibilities to explore future features such as PEPPOL compliance and inbox triage capabilities with FUJIFILM Process Automation.

“Health NZ began auto-posting invoices from day one, which significantly reduced manual tasks. This project demonstrates how the Finance and Technology teams are using AI to minimise back-office costs, allowing more resources to be allocated to frontline services,” she says.

 

If you have any questions re the above feature article, please contact the editor Rebecca McBeth.



Read more FEATURES


Return to    eHealthNews.nz home page