How Affinda Group is Driving AI Transformation
Backgrounds in quantum physics and investment banking helped two brothers build a world-leading AI company. Co-founder, CEO and one of the siblings, Tim Toner, explains how.
Fact file
Co-founders: Dr Ben Toner (left), chief scientist, 45, Tim Toner (right), CEO, 43
Investors: High-net-worth investors, family and staff members
First customer: Deloitte, 2018
Headquarters: Melbourne, plus staff in 10 other countries
Staff: 70 and growing
What’s your elevator pitch?
“Affinda Group is an Australian-founded company revolutionising document processing through AI-driven automation. Our technology extracts actionable data from any document type across 56 languages and is relevant wherever there’s a document-intensive workflow with people doing lots of manual tasks.”
How did you get it off the ground?
“Ben and I have complementary skill sets – his in quantum physics and mine in investment banking. In 2012, we didn’t realise how valuable that technical and commercial combination would be for technology that’s driving AI transformation. You not only need to build the technology but also work out where to point the products in terms of the investment case for our customers. It’s that combination that led us to these high-value and complex business problems for large organisations.”
What was the problem you were aiming to solve?
“Our first product, Draftable, is document-comparison software and its predominant use is to compare legal contracts. Then came Affinda, which is document-processing software for use by every industry across any document type. The Affinda technology leveraged our understanding of documents from Draftable but took us much longer to build – it was hard to teach a computer to read a document in the way that a human would. Both products enable businesses to automate high-volume workflows, reduce errors, drive cost savings and improve productivity. Draftable is more of an off-the-shelf, standard SaaS [software as a service] product – the free online version was used to compare documents more than five million times on our website last year. Affinda is more complex with a higher level of customisation – we design a solution that embeds our technology into the workflow to optimise the document processing. Today we have 1500-plus customers in 80 countries.”
How does it work?
“Affinda and Draftable are at the less sexy end of AI but we celebrate that. Documents are pervasive across pretty much every company and industry, with a lot of people working on the data trapped in them. Employment website SEEK is a customer and Affinda extracts information from CVs to match candidates to job descriptions. Another example is loan settlement automation, where we’ve built a verification engine that validates, say, 300 different rules, checks and balances. It serves up the exceptions to the humans so you need far fewer people and takes the settlement time from three days to between 10 and 30 seconds. It shifts people to higher-value work or lets companies handle five to 10 times the volume with the same number of staff. Parts of it draw on self-driving car technology. On the road, it’s ‘That’s a pedestrian’ or ‘That’s a stop sign.’ In our world, it’s ‘That’s a header’, ‘That’s an image’, ‘This information is connected to this.’”
How did you convince investors and customers?
“Ben’s business model was that there’s a global demand for these products so if we build great products and make them available online to let customers trial them, our business will grow. I thought it sounded too easy but that’s exactly what’s happened – 95 per cent of our paying clients found us online and tested the product. We only hired our first salesperson in 2019.”
What’s next?
“We’ve raised a lot of money – $22.9 million. We’re scaling quickly and will keep building complementary products while being focused on creating the best businesses we can, making our team proud and customers happy. This is our life’s work so we’re 12 years into a much longer journey.”
