Betting the Farm on AI at Fonterra: OpenText Sydney Summit 2025

In a world drowning in corporate data, artificial intelligence is emerging as the life raft that major organizations are desperately reaching for. This was the clear message from OpenText's February 20 Sydney Summit, where industry leaders shared their struggles with information overload and their hopes for AI-driven solutions.
The scale of the challenge is staggering. Take Fonterra, the New Zealand-based dairy giant, where employees have created more than 29,000 SharePoint sites – averaging more than one per person for the 20,000-strong workforce. "This is not sustainable from a data footprint perspective, and from ensuring that integrity of that information," warns Caroline Wishart, Fonterra's Head of Information Management, Data and AI.
For Fonterra, a century-old dairy cooperative, the stakes are particularly high. The company is betting big on AI, with expectations of using Data and AI as a significant contributor to achieving their strategic ambitions.
Their vision extends to supporting a transformational shift – exploring solutions that support their Research and Maintenance teams to access, analyse and generate insight from their data information and IP.
“The key to make progress going forward in the information management space is to automate our approach to deduplication and the protection of the right levels of information,” said Wishart.
“One of the key priorities from my role as head of information management is ensuring that we are taking care of our information that we've historically gathered over the 100 years that we've been in business as a dairy cooperative, and then ensuring information is fit for the future as well.”
OpenText Extended ECM (xECM) is currently used by about 4,000 Fonterra asset managers in manufacturing sites. It enables asset maintenance staff to access the very latest asset information through sap.
“In the last year or so, we rolled out some of our data classification products, particularly across Microsoft so labelling is now mandatory for emails. We are getting people to think about protecting the information they are storing and sharing.
“Another focus is retention, in terms of trying to fix the past to enable AI to ensure its accessing current, accurate, and suitable information.
“It’s also reducing duplication where there’s copies in SharePoint and copies in the shared drive.
Fonterra's Head of Information Management, Data and AI, Caroline Wishart.
“Our three priorities are ownership of information, retention, so only retaining and content as long as we've made it, and classification of data sets.
The company’s manufacturing business unit has identified AI opportunities with an expectation of contributing to their goal of significant cost out in 2030.
“We’re also working on how we can utilise AI for our Policy and Procedure Library which is stored on Content Manager. In Palmerston North our research centre has many staff including about 100 PhDs focused on new products, research and technology.
“We have been working with them on a solution called R&D driver, to enable our scientists and researchers to look across all the repositories and all the research that has been generated by us as well as ingesting the vast amount of international research.”
In his keynote presentation to the Sydney Summit, Mark Barrenechea, OpenText CEO and CTO, emphasised the importance the company is placing on its Aviator AI platform.
“Information is our platform. Whether you're on Bedrock in AWS, or you're in Vertex within Google, or you're in co-pilot in Microsoft, or you brought your own language models, our Aviator Studio will coordinate all of it for you.
“It’s obvious to us that Business AI is relevant in every industry.
OpenText CEO and CTO, Mark Barrenechea.
“We are going to embed it everywhere. It's going to be just like you have a search button, you’re going to have our Aviator button embedded everywhere. It's going to be just part of what we do. We have 15 Aviators and over the 100 agents and agentic AI is the next big step.
“In April, we're going to unveil our first Agentic AI. We now have our agents talking to agents to complete workflows and, multi-step complex processes where no human is involved to actually make decisions and learn from them.
“We started a few years ago with machine learning and Predictive Analytics.
“We then stepped up and started to support Gen AI tools. We started with open source and we’ve now moved more formally with deep support into Google and Vertex. We've now added what our security platform CoPilot support and we'll have a few big announcements in April on how we're expanding this even further,” said Barrenechea.
The Power of AI
An ECM Specialist at one of Australia’s largest power generation and transmission companies spoke at the Summit of her excitement about the possibilities for integrating AI.
Some of the company’s oldest facilities, which are spread all across the country, date back to the 1960s, meaning there is over 60 years of content stored in OpenText ECM content server and Extended ECM (xECM) for Engineering.
“We're really keen to have the ability for a user, for instance, to ask which vendor has provided valves on an asset over its lifetime. Being able to find those results accurately would be a huge timesaver,” she said,
The firm’s core ECM focus is controlled engineering documentation, including technical drawings, specifications, manuals and operating procedures.
“One of our biggest challenges with engineering documents is making sure that our users are using the latest revisions necessary. When dealing with complex infrastructure at the scale that we do it’s not just an inconvenience, if users are using later revisions it’s actually a really serious safety risk.
“We rely heavily on the transmitter module to share documents with our external parties. We make sure that it maintains full traceability, so this is anything for sending specifications to our vendors, updated process flow diagrams to our regulators. We need to know who received those documents, what they received and when they received them.
“It's extremely critical from an auditing perspective,” she said.
OpenText was initially deployed by APA in 2018. A concerted push to increase usage commenced in 2023 which has now surpassed one million controlled documents in the ECM and 2.5 million objects.
NSW Office of the Director of Public Prosecutions
An OpenText Content Manager user, the NSW Office of the Director of Public Prosecutions (ODPP) has developed its strategy in the past financial year and is currently running a couple of AI Proof of Concepts.
Daniel Campisi, Manager, Applications & Information Management, NSW Office of the ODPP, said, “We have a number of challenges in the legal space. There are challenges with AI being involved in in the development and the production of legal material, so we, we will need to make sure that the AI will be creating accurate information with no hallucinations and making up cases.
“AI will generate documents, we just need to make sure that they are accurate and validated by humans in the middle.
“We have to deal with facts, the actual details that have occurred, so it's very difficult to reimagine the information. So, in our organisation, we will try to reimagine how we can make the information available to our staff.”
“We're doing that through integration and information. So, the complete story is known, and it will be easier and faster for our staff to understand the complete story and not a fragmented one.”
NSW Office of the ODPP Manager, Applications & Information Management, Daniel Campisi.
Campisi has overseen the replacement of a legacy matter management system at the ODPP with a new Web-based application built on the Appian low code platform, and with the assistance of solution provider iCognition, integration with a new instance of Content Manager.
“With the assistance of iCognition we leveraged a lot of the out-of-the-box functionalities of Content Manager, without people using Content Manager directly or this application being installed on their desktop.”
“iCognition’s RM Workflow allows people to edit and save a document directly into CM within the staff doing anything other than saving.”
“We are trying to ensure that the default behaviour is that people save documents back into the system. Currently, all of the NSW State Records record-keeping requirements are satisfied with minimal and often no, actual interaction.”
The Information Management Lead at an Australian state electricity transmission network is looking forward to the potential of OpenText Aviator.
“The power that AI will give to our OpenText Platform is enormous in terms of intelligence and governance,” she said.
The company has over 600 staff and has deployed OpenText Extended ECM (xECM) with Enterprise Connect on premise.
“OpenText was already installed when I arrived but it was a challenge to use and a lot of workers resorted to workarounds which diminished the system's usability” she said.
“We also had information stored in lot of different locations, on Servers on SharePoint and in OneDrive.
“We undertook a big push to make the ECM system more usable and started a manual migration of data in May 2024 now nearing 800,000 items. We are migrating 5000 items a week and have probably got 6-12 months left of this journey before migration is complete.
“The benefits include reduced administrative overhead and it also fosters collaboration when information lives in one location and everyone knows where to get it.
“We are now about to release OpenText Mobile to provide workflow approval for anybody on the road and give accessibility to field staff so they are always accessing current information.”
“Customers who have been on a journey with us and consolidated and managed their information in more singular ways, are in a much better position today, as they start to unlock the value of that data through new tools, such as AI,” said OpenText’s Barrenechea.
“Whether you're a Salesforce shop, ServiceNow, Google, Microsoft, Azure or Amazon, Oracle or sap. We always come in underneath it and make it all work together.
“Information is our platform. We make all those apps and multi-cloud work we're going to do the same thing in the AI world.”
Bill Dawes is the publisher/editor of Information & Data Manager (IDM)