The health fund set out to modernise the way everyday claims move through the business—so members notice quicker turnaround times and clearer communication, and staff can spend more time talking with people rather than keying in data. Built on Microsoft, the new human‑in‑the‑loop platform turns a high‑volume, manual process into a streamlined flow that lifts quality today and sets the fund up for the future.
A better rhythm for claims processing.
As a long‑established industry health fund, the organisation has a strong mission to provide affordable health cover and now supports hundreds of thousands of members across Australia and New Zealand, including its Auckland‑based members. With that scale comes a daily challenge: thousands of claims need to be reviewed, validated and progressed with care—particularly ancillary claims, which make up most of the volume and have traditionally required significant manual data entry and follow‑up.
The fund had already been looking at how artificial intelligence could be used responsibly to improve experiences for both members and staff. Its roadmap identified claims as a priority, with simple ancillary claims offering the quickest way to reduce pressure while maintaining decision quality. Earlier optical character recognition (OCR) trials delivered mixed results—effective in some cases, but unreliable with the varied formats members submit and the complexity of provider codes and service descriptions. The conclusion was clear: any new solution needed stronger document understanding, combined with a clear role for people to stay involved, ensuring accuracy and building trust.
This wasn't just about efficiency. It was about confidence—within the organisation and with members—that a technology‑enabled claims journey could still feel personal. That meant cutting down double handling, identifying missing information earlier, and freeing up staff to focus on claims that genuinely need expertise, while also laying the groundwork for improvements in contact‑centre knowledge and, over time, stronger fraud detection.
Using the Microsoft platform to drive flow
The fund chose to build where it already invests—on Microsoft. The minimum viable product (MVP) for simple ancillary claims uses Microsoft Azure to bring together document intelligence, workflow integration and a human‑in‑the‑loop experience, allowing automation to handle the routine work while people apply judgement where it counts.
At the start of the process, Azure AI Document Intelligence reads invoices and receipts, including handwritten details, extracting key information to reduce re‑keying and improve first‑pass accuracy. Confidence scores are built in. Where scores fall below a set threshold—or business rules require a second look—the claim is sent to an easy‑to‑use review screen so assessors can quickly check, correct and move on. The result is faster processing without giving up control, helping build trust and deliver value sooner.
The solution fits neatly into existing ways of working. It connects to the current workflow system for intake and exceptions and passes validated data to the existing decision engine—respecting established rules and governance while removing unnecessary friction. The Phase 1 goal is clear: achieve straight‑through processing for at least 40% of ancillary claims, with scope to increase coverage and capability over time. Targets and acceptance criteria are clearly defined, and the solution is built within the fund's own Azure environment to support long‑term ownership.
Because this is about transformation, not just new tools, the technology choices support strong operational discipline. Azure App Service (supported by Azure Logic Apps and Functions where needed) manages the flow, Application Insights provides real‑time visibility of performance, and modern security controls protect data by design. It's a practical, focused approach—using a small number of Microsoft services, selected for purpose and integrated to deliver clear business outcomes.
From faster claims to a smarter organisation
The first thing that changes is how the work feels. Members benefit from quicker, clearer experiences—documents are understood correctly more often, missing details are identified earlier, and claim status is easier to track. For staff, less time is spent on re‑keying and more on claims that need judgement. Given the volume of ancillary claims, moving routine cases through smoothly frees up capacity for complex matters and higher‑value member interactions.
The platform also strengthens the organisation's resilience. By running it within its own Microsoft environment, the fund keeps control of its data, aligns with its security standards, and reduces the effort required to support and enhance the solution over time. The human‑in‑the‑loop approach isn't a compromise—it's a core design principle that builds trust and provides ongoing feedback to improve extraction accuracy, refine rules, and safely expand to other claim types.
Beyond claims, this work opens the door to further improvements. Structured data from document processing can feed into fraud analytics over time, adding richer signals alongside internal and industry data. At the same time, the fund is exploring Microsoft‑based knowledge management for contact‑centre teams—supporting a "one place to ask, clear guidance to act" approach that shortens time to competency and improves consistency for members.
People remain central to the change. The programme includes clear acceptance criteria, governance, and a deliberate build‑and‑transfer approach so the fund's own teams can run, refine and extend the platform themselves. Starting with a well‑defined scope—simple ancillary claims—delivers early wins while building capability to scale. As accuracy improves, quality assurance effort reduces, and straight‑through processing increases by design. The outcome is a measured path to enterprise readiness, driven by results and underpinned by the fund's growing Microsoft capability.
The result is more than faster claims. It's a smarter operation—one that turns variation into insight, balances automation with human judgement, and uses the Microsoft platform as a long‑term foundation for service excellence. For a member‑owned fund, that matters. It means fewer delays, clearer answers, and staff focusing their time where empathy and expertise make the biggest difference—while maintaining a strong platform for continued improvement on the fund's own terms.
