The Australian Human Rights Commission (AHRC) performs a critically important role in society – and is now also a global pioneer in the use of artificial intelligence and machine learning.
A project, to automatically classify records and documents – the first of its kind in Australia – has already attracted the attention of the National Archives of Australia which has branded the Commission as a thought leader in electronic document and record management. It has identified the solution as a reference point for other government agencies which grapple with similar document management challenges.
The AHRC transitioned many of its core applications to Microsoft Azure and adopted Office 365 over three years ago, following that with a Dynamics 365 deployment, recognising the cloud would deliver the foundations for ongoing transformation, operational improvement and enduring value.
Now, working with independent software vendor RecordPoint and its federated data management platform called Records365 the Commission has built an AI-infused document and records management solution in SharePoint Online.
According to AHRC CIO Ron McLay, the migration to Microsoft cloud has been transformational for the Commission and delivered high resilience and security, as well as the foundations for continuous improvement.
“Moving to the cloud meant that there was more resilience to the network. It has enhanced our security as well. It seems to me that many people have an idea that moving to the cloud means that you’re sacrificing your security, sacrificing the integrity of your network. But it’s actually the contrary that’s true. Microsoft as an organisation has a much greater ability to protect against zero-day threats and against substantial attempted penetrations and malicious software and phishing attacks.
“We haven’t had a security incident since we’ve moved to the cloud.”
What the AHRC has also had is increased flexibility for personnel who are able to securely access applications and information when and where they need it. With robust foundations in place McLay and his team have also been able to accelerate the pace of innovation and transformation.
McLay says that Project RADICAL (Record And Document Innovation and Capture – Artificial Learning) is one of its most exciting recent developments at the Commission and tackles a significant challenge that had arisen.
Previously the AHRC had relied heavily on paper documents and electronic copies to allow its million documents to be shared or searched. Ryan McConville, AHRC’s Information Manager, stated in regard to those electronic records, “We were using a flat file structure prior to that, and we had all sorts of legacy problems with complex nested folder structures,”.
“Perhaps the most exciting development in recent times, is that we’ve been able to leverage from SharePoint Online, and we’ve used a product called Records365 from RecordPoint, which is a federated data management platform that leverages the power of machine learning to assist organizations in intelligently applying compliance policies to unstructured content. Records365 was able to learn what the different classes of information within AHRC look like and then make policy recommendations based on how similar a document is to existing documents managed by Records365.
McConville remarked, “We were their first customer worldwide. We’ve been working with them now for over a year on the machine learning elements of their product, Records 365, and that’s been really exciting, that we’ve built this environment in SharePoint Online. We’re midstream migrating the units across to that environment, but we’ve been in production now for eight months.”
Records365 uses advanced machine learning techniques to classify electronic documents, reducing the risk of human error, while increasing the volume, accuracy and consistency of records classification.
After an initial training period, accuracy of 80 per cent plus is being achieved at the AHRC with further improvements anticipated over time.
The simple user interface has driven high user uptake and incorporates browsing and search capability helping to deliver a 5 per cent productivity boost. The Commission also expects to be able to significantly cut the cost of physical document storage – which had been costing $17,000 a year.
RecordPoint worked closely with the AHRC, using project methodology that adhered to the National Australian Archives’ (NAA) recommended approach to implementing an EDRMS and following an in-depth assessment of the AHRC’s requirements. A comprehensive program of work and change management plan was developed to ensure its success with the solution delivered cost effectively – a more traditional EDRMS without AI could have cost three to four times as much.
According to Anthony Woodward, Chief Technology Officer, RecordPoint; “The successes seen with the Commission’s project RADICAL demonstrates to other government agencies that AI/Machine learning is a viable, affordable option for automating records management in the public sector. This is significant, as historically the high costs of implementing AI-based systems has been beyond the reach of most government agencies.”
The success at AHRC has already attracted the attention of other agencies, says McLay.
“It’s so futuristic from the federal government’s point of view that there’s been many agencies that have come to see demonstrations of the environment we’ve set up in SharePoint Online, and Records365. We have probably run demos for maybe two dozen or even more agencies now.”
Although the AHRC has about a million documents to manage – compared to other agencies that is quite modest. McLay says he knows of peers who are tasked to save 13,000 documents each week.
He believes that other agencies struggling with flat file structures and manual management, and grappling with versioning and duplicates could benefit from a similar approach.
“As we’re moving everything into RADICAL, it’s much better organised. We’re eliminating all the duplicates. We’ve introduced versioning to our staff now, so staff are spending a lot less time looking for documents or wondering about where they should save them.
“What we’ve done with our iteration of SharePoint alone is we’ve set the environment up so there’s no mandatory metadata. It’s all captured through the way we’ve set SharePoint up. We’ve avoided any third-party add-ons.”
“We’ve avoided any customisations, and we’ve been very disciplined. We’ve kept it as simple as possible so that the user experience is crisp and clean and easy.”
The AI infused Records365 federated data management platform is an important innovation – but far from the last on McLay’s roadmap.
“I’m particularly interested in a little side project that we’ve been working with Microsoft on for some months, and that is a translation engine that can translate documents into easy English – a subset of the English language, 850 words
“It also includes pictures, and it’s targeted to people who perhaps have English as a second language, or people whose comprehension isn’t able to deal with the full English language itself, but a subset. That’s exciting, because that’s helping the disabled community, and making our information more accessible to anyone who needs the services in the Human Rights Commission,” which is clearly a part of its charter.
Because the AHRC engages extensively with the disabled community McLay is always on the lookout for solutions that make sense for these stakeholders. He has for example used Skype Broadcast to great effect.
“We used it, actually, when we were demoing Project RADICAL to about 10 agencies in this particular presentation. The people that dialled into that session, they were seeing a transcription, a live transcription, of the presentation. That was a great feature to be able to roll out.
“We want to be consuming cognitive services wherever and whenever we can find opportunities to do it. I think there are a lot of benefits for people who are visually impaired, people who have hearing issues, people who have comprehension issues.”
McLay can see a role also for chat bots. “The way I would see it operating is perhaps in the future, hopefully not the distant future, that we would have clients calling into the Commission, and their query would be answered by the digital bots who perhaps could swap into the language of the caller as required. I see that the ATO is already claiming that their digital bot is answering 90 per cent of first-time inquiries, which is quite a phenomenal rate.
“So there’s definitely a big future in machine learning and artificial intelligence, and we’re already in the cloud, so it’s there, ready for us to use.”