Does the hyper rapid rise of artificial intelligence (AI) signal the end of the world?
Or the start of a new era?
Jason Verlen, senior vice president, product management, at CCC Intelligent Solutions Inc., a cloud platform for the insurance and automotive industries, tells PYMNTS that, in his view, it is unequivocally the latter — we are stepping into an era with tremendous potential.
“This really is an exciting time,” Verlen said, adding that the business landscape is starting to leave behind the “Wave One” of AI platforms pre-trained on public data and beginning to enter what he refers to as “Wave Two.”
“Wave Two is where organizations take control, using generative AI on their own in-house data to solve narrower problems with domain-specific, secure, and high-provenance data,” he explained.
What’s more, by using wholly-owned company data to power internal AI engines, businesses can leverage the technology’s generative capabilities to come up with insights and advantages unique to them that are impossible for competitors to replicate.
Read More: Generative vs Predictive AI’s Role Across the Future of Payments
Data-rich environments lead to data-driven transformations
The multi-decade digitization of the business landscape has produced an operational enterprise infrastructure that now leaves savvy organizations ready to gather, process and scale insights via a future-fit flywheel of data powered — and frequently automated — solutions.
Verlen says that he sees two key things across the payments domain where generative AI can have an immediate impact by building upon the transformations predictive AI has already wrought.
The first, he says, is generative AI’s capability for evolving fraud detection strategies and approaches.
“AI can create synthetic data, allowing you to train anti-fraud models without having to resort to labor-intensive [legacy approaches],” Verlen noted.
The challenge with fraud is the rarity of actual positives in the data to train from — that’s where creating synthetic data can be beneficial. Because large language models can look at a lot more data than previous approaches, they may be better able to detect certain patterns with supervised learning that previous generations could not.
Last but not least, Verlen added, is that by taking advantage of what many view as a weakness of generative AI — its tendency to “hallucinate” or create false content and unsourced fabrications — firms can even potentially get the jump on bad actors by charting out and creating new attack vector possibility maps.
“There are certain domains where going off the rails a little bit is actually a positive, where it can give you a look at other possible scenarios you may not have come up with yourself,” explained Verlen. “A lot of fraud detection is done by looking for outliers — which is frequently a labor-intensive and expensive process. Hallucinations essentially can show you vectors that aren’t there yet but may be coming.”
Read also: AI Regulations Need to Target Data Provenance and Protect Privacy
Taking what’s been done already and pushing it forward
The predictive AI solutions and machine learning (ML) tools of the past few decades have already accelerated the agility and effectiveness of many business processes.
For generative AI to be successfully commercialized, it will have to go even further.
“A lot of what’s being done with large language models (LLMs) can actually go much further — not just improving performance, but unlocking new business use-cases [and even new businesses],” said Verlen. “The possibilities inherent in AI for driving new horizons of efficiency and productivity in the macro environment are almost unbelievable.”
Already within CCC’s industry, AI is being used across various workflows and use cases to augment historical processes and, in certain cases, entirely transform them.
Verlen gave an example of a car accident where the vehicle is totaled and determined to be a “total loss.”
In this case, Verlen explained, an AI tool can look at vehicle photos and identify the damage, allowing the insurance carrier to determine if a vehicle will be totaled. If so, the insurer can reach out to the lender on that vehicle and make a payment to get the title for the car so it can be salvaged. And all of this takes place in as little as a few hours rather than the days or weeks it previously took to process the transaction.
“AI will greatly increase the speed and convenience of payment processes — and it’s not just the payment workflow [where AI will have an impact], but across everything that goes around the deployment of the payment by creating the specific context that drives a response,” he said.
Still, Verlen cautioned that “these generations and waves of technology always coexist, and they will continue to for a while,” with manual processes being augmented by AI solutions and vice-versa.
He emphasizes that this period of “coexistence” will create new business value through a “collaboration with generative AI and human brainpower” to create something deeper and more meaningful because individuals will have “more time” to look at the things that are actually unique.
“Every time we have a pivotal moment, it’s projected to be a doom loop for humanity in terms of employment — and what happens 100% of the time? The complete opposite. There is an explosion of productivity, more jobs are created, and we enjoy a more vibrant economy,” Verlen said, adding, “when I look at this [generative AI] technology, the term ‘replacing me’ doesn’t accurately capture the nature of the transformation.”
Rather, he says, AI will serve to augment labor-intensive workflows by making them smarter, more productive, and increasingly optimized.
We’re always on the lookout for opportunities to partner with innovators and disruptors.
Learn More