Canberra-founded company Xaana leads on artificial intelligence
Canberra at the forefront for developing artificial intelligence (AI) based on collaborative intelligence (CINTEL)? Absolutely. And Xaana is also working to professionalise data science and build greater awareness around AI.
Canberra-founded company Xaana works with companies and government departments in Australia and globally to help improve business processes using AI. Xaana has recently launched the world’s first AI twin. And it’s looking to professionalise data science. Editor Serina Bird chatted with Xaana partner Matt Andrews to find out more.
Canberra HQ
I’ve walked past the Xaana headquarters on Mort St for months on my way to and from the light rail stop in Civic and wondered what the high-tech-looking company was all about. So I was intrigued to find out more. Then, I was contacted about the release of Zebra, touted as the world’s first AI twin.
Xaana was founded by Dan Saldi, whose impressive CV includes graduating with applied artificial intelligence from the Massachusetts Institute of Technology. After founding Dexus One and then selling it to Deloitte, he worked as a Deloitte partner before taking the plunge and starting Xaana.
Xaana creates software products for integrated human-driven and machine-assisted analysis. In particular, it does large data integration work with enterprise resource planning (ERP) systems and then adding generative AI to it. In other words, it helps bring information from various programs that a large company might be running to bring them together and then use natural language to pull up relevant information from large data. For instance, financial data can be put in a report along with employment or other data in a secure way and then queried together, or someone can use Xaana Generative AI to make the report on demand. This saves staff from having to pull data together manually from different systems.
Xaana’s clients include Australian Government departments and large companies such as Shell Petroleum, Kmart and Bunnings. Its 50+ staff are mostly in Canberra, with some working remotely. It’s an under-told Canberra success story.
One of the reasons Xaana has chosen to base itself in Canberra is because it’s more relationship-driven than other cities or towns. ‘There’s three degrees of separation in Canberra, so those networks are really powerful,’ said Andrews. ‘And there are some really high-quality universities here that have fantastic students.’
Human in the loop
I’ve been watching The Terminator film series with my sons, and that’s led to discussions about the possibility of machines taking over the world. Back in 1984, when the first movie came out, that seemed far-fetched, like something created by the overactive imagination of a script writer and an excuse for a cold-war nuclear plotline with car chases and a big bulky hero (i.e. Arnold Schwarzenegger).
But forty years later, in 2024, the possibility of machines to learn from humans and take over our roles – even if not to destroy us – is real. Who in the tech, startup and entrepreneurship space hasn’t talked about ChatGPT and what it means for jobs of the future?
Xaana under Saldi is at the forefront of developing AI, but its focus is on collaborative intelligence (CINTEL), or ‘human in the loop’. CINTEL is about ensuring that machines aren’t making decisions; humans are. But humans can harness AI to create efficiencies.
According to Andrews, many people work in government departments where there are disparate systems holding pieces of information. If a report is needed, someone needs to view information, compile it, sort through it, and then write it manually.
‘What we see AI doing, and what we have been doing, is it will automatically always pull all this information in. And then you can ask a particular question in natural language,’ he said. ‘The big difference from things like ChatGPT is it’s not scouring the internet. You ask it a specific question, such as “Can you please tell me the current cost of administering this particular program?”. It will give you that information with citations from your own database. This means that the human is making the decision to compile the report and can then validate the information using the citations.’
Andrews doesn’t believe we need to be scared of AI technology taking our jobs or taking over. ‘AI enables people to focus more on the things that we as humans are innately better skilled to do than AI in terms of assessing things, making decisions, and using this wonderful thing called human judgement that no AI can ever do,’ he said.
‘AI can place together things in a logical sequence based on what’s happened before. But it doesn’t necessarily see the entire picture. And it doesn’t factor in things like emotion, nor that element of human creativity, judgement and decision-making. Not does it understand compassion.’
Zebra – the AI twin
In October 2023, Xaana launched Zebra. Zebra seeks to address the limitations of using open-source AI technologies such as ChatGPT in secure environments such as government and the military. Zebra consists of a suite of products Saldi has developed, including large data integration, supply chain automation and generative AI, which all use CINTEL concepts.
For instance, Zebra can sit over Microsoft Outlook on a secure system and pick up every invoice emailed in. It can check each invoice's ABN and account details and whether the organisation is already on file. Zebra will realise if there have been any changes to the account details on file, such as a different bank account. It will then notify someone to contact the vendor and ask if bank account information has changed, thus preventing possible spam or fraud attempts.
Andrews noted that Zebra has potential military applications as Zebra can bring together structured and unstructured data. For instance, it could compile information from social media and sensor data from aircraft and ground vehicles for front-line decision-makers.
Data science as a profession
What is data science? Well, it’s broader than coding. Andrews explained that coding is a specialised area within broader data science in the same way you could say that machine learning is a small, specialised area of broader artificial intelligence.
But there’s no legal definition of data science, no professional body, nor formal requirements in terms of who can call themselves a data science. This is something that Xaana wants to change.
Andrews has come from working with CPA Australia and is familiar with the process of setting and retaining professional standards. He met founder Saldi after Saldi presented on some popular CPA Australia webinars. Saldi then invited Andrews to work for Xaana and to help professionalise the data science profession.
‘We want to set up a professional body that mimics what you’d get with say a CPA for data science, where you have an entry requirement, a work experience requirement and an education requirement before you will say you are a professional,’ said Andrews.
Xaana is in discussions with Australian universities to deliver the program's academic component and with industry and software companies about the work experience component. As data science is a rapidly evolving sector, a continued professional development requirement will be key. ‘The amount of change that we’re seeing in even 12 or 24 months is phenomenal,’ said Andrews.