Kiran Ratnapu, founder and CEO of Deep Relevance, says that using AI for fraud detection provides faster results at lower cost than conventional rules-based approaches:
When we created Deep Relevance, we set out to identify fraud faster and cheaper than had been possible in the past. Traditional rules-based systems were missing fraud that we knew AI could find.
The AI models developed by Deep Relevance are able to identify and assess risk signals in an organization's spend data, such as when an employee transposes digits in expense claim submissions. As Donna Wilczek, VP, Product Strategy and Innovation at Coupa, told me today:
The core is defining the tendencies of what a person will do that is committing a fraudulent act.
The software then creates alerts based on this risk analysis, prioritized based on value at risk so that internal auditors or finance personnel can focus on the most significant. Areas of fraud detection covered by the technology include conflict of interest, bidding integrity, sham vendors, fraudulent invoices, inflated expense claims, duplicate expenses, and personal expenses.
The acquisition now allows Coupa to apply this AI technology to its vast database of spend data collected from its customer community, as CEO Rob Bernshteyn explained in the opening keynote of its Coupa Inspire EMEA conference in London this morning:
Their concept was to do it company-by-company. Our vision is to take their capabilities and work with them hand-in-hand to apply this artificial intelligence to hundreds of billions of dollars of spending.
The AI will also be able to look for patterns of data and behavior across expenses, purchase orders and invoices within the Coupa system. Most fraud detection systems work within a single category of spend, says Wilczek, so the ability to look across different types of spend extends the reach of the technology:
Someone may never trigger an individual alert, but their behavior does through holistic view.
Coupa made a significant investment in AI at the end of last year with the acquisition of Spend360, a world leader in the use of deep learning for spend classification. That technology has been applied to Coupa's spend data and powers its Risk Aware supplier risk management tool, which is now generally available following a period in early release. During that time, there have been a number of cases where customers have been able to use the community benchmark to identify and adjust supplier risk exposure.
As well as using Coupa's own data store, Risk Aware also uses third-party data collated by technology acquired with data aggregation platform Riskopy earlier the year. The third AI-related acquisition was Trade Extensions, a strategic sourcing and advanced optimization platform that’s used by many of the world’s largest brands to run complex sourcing for categories such as fresh produce or direct materials. This brings Coupa into the direct procurement market.
Wilczek says the cultural fit when making an acquisition is as important to Coupa as the technology. The Deep Relevance team will be working closely with Spend360's data scientists and AI experts as well as other parts of the Coupa organization to integrate its technology into the product architecture.
Coupa also today announced general availability for US customers of its integration of Amazon Business into its procurement catalog. For customers that opt in, this gives users access to products in the Amazon Business marketplace from within Coupa, ensuring that any purchases conform to the organization's procurement processes as defined in Coupa.
Another astute acquisition of some very useful AI technology and skills on Coupa's part. Putting this together with its large and fast-growing store of spend data allows the vendor to offer its customers significant practical value from the application of AI.