Pinpointing specific information in massive swaths of financial services contractual documents was the driver that created a need for KPMG International Ltd.’s new Ignite artificial intelligence ecosystem, says a software engineering boss there.

“Tons of contracts” and documents existed at the professional services network that were full of valuable information, but there wasn’t any way of extracting the good bits, according to Kevin Martelli (pictured), principal of software engineering at KPMG.

“It really evolved into a whole end-to-end complete platform,” he said. Interestingly, KPMG has also just gotten a patent for it.

Martelli spoke with John Furrier, host of theCUBE, SiliconANGLE Media’s livestreaming studio, during the recent Red Hat Summit. They discussed the Insights tool KPMG has developed. (* Disclosure below.)

Containerization and analytics

Containerization is a fundamental element of the AI platform. It’s all about deferring risk — some documents are sensitive: certain workloads can be run on-premises via OpenShift containerization, Martelli explained, whereas other, less sensitive material can be “burst” over to another cloud, including public ones.

“So you can get advantage of their scale, their capacity,” he said.

Red Hat developed the OpenShift containerization-underlay that KPMG has chosen as an option for deploying its AI platform. It’s “having the ability to move workloads quickly and fast where they make sense,” Martelli said. “Where the security and risk is aligned is something that would make a successful use case. Application users of Ignite — the developers, the data scientists, the business users, the analysts — they all need to interact with the platform.”

That’s led to it being developed with a conscious differentiation between the underlying Red Hat infrastructure, with the users “getting the insights about getting the efficiencies in the platform.” The background mechanics are separated from usability; in other words: the movement of workloads, scalability and so on is handled unintrusively by the Red Hat element.

Artificial intelligence and data literacy

Introducing into production the more complex AI algorithms needed in products like Ignite is tricky for the industry, though, according to Martelli. KPMG has performed studies which illustrate that although businesses want to adopt AI and ML, “industry leaders are a little bit concerned about how fast this adoption is going. And is it going too fast?” Martelli asked.

That’s because digitization accelerated rapidly during the peak of the COVID period — and continues to. Martelli reckons he’s seeing organizations now, post-pandemic peak, producing applications quicker while adhering to earlier, existing control frameworks even.

Outstanding issues include ethics, governing process, security and privacy. Guidance, however, may come quite soon as the Biden administration, although open to AI, could introduce regulations around the technology. That might well help corral some of those challenges, Martelli believes. That guidance from government could make the questions around AI clearer to solve, in other words.

Machine learning biases over time is one example of something that can’t be allowed to creep into the algorithm deployment, according to Martelli. “There are still struggles around that,” he said.

Challenges don’t stop there, though. Everyone should be involved in data, according to Martelli. By that, he’s referring to data literacy throughout an enterprise.

“Really having our folks understand the data and empowering them to be successful” has been hard to achieve, he explained. “I need some analysts to use it; I needed some data scientists and engineers, but what we’re starting to see now is there’s larger programs across the board where it’s more holistic at an organizational level,” he says of the direction things are going.

Watch the complete video interview below, and be sure to check out more of SiliconANGLE’s and theCUBE’s coverage of Red Hat Summit. (* Disclosure: TheCUBE is a paid media partner for Red Hat Summit. Neither Red Hat Inc., the sponsor for theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE