Next time you call your doctor or your financial adviser, the voice on the other end of the line may sound like them but not be human.
Rapid advances in generative artificial intelligence (‘Gen AI’) have created robot voices indistinguishable from familiar human vocal patterns. And while businesses have used computer voices for decades to serve customers and answer phones, their unsettling staccato delivery and lack of empathetic responses marked them as artificial.
But when backed by the recent explosion of AI algorithms, rapid processing, deep data retrieval and ubiquity of the cloud, these emergent systems pass the ‘Turing Test’ or ‘Imitation Game’ for computers that exhibit human-like intelligence.
Brett Barton, Unisys’ newly minted global AI practice leader, illustrates this point: “In my past [role, before working at Unisys], we leveraged AI — including using actual voice of call centre operators, in this case financial advisers — to answer the phone and then the live person steps in,” Barton said.
“The [customer] never has any idea that they were talking with anyone but the financial adviser.”
Landing in Australia during the recent global CrowdStrike blackout, Barton spoke with CRN Australia about how AI is seeping into business, from customer-facing functions to the back-office.
In addition to talking about the use of AI for voice-based customer service, he also touched on its use in healthcare billing.
Barton said AI could help smooth whole-of-patient care from doctor’s surgery to practice manager’s office.
“I’ve done work in the healthcare space where we use AI to help schedule and essentially bill insurance companies for multiple comorbidities for a patient in a single visit,” Barton said.
“In order for the physician to get that pay they have to address each of those comorbidities. And in order to be able to knowledgeably treat those, tests have to be run, and it has to be run prior to that visit with the physician with enough time for the results to be generated and reviewed by the physician. And so ultimately, it's a space where AI can bring a scale of efficiency.
“In the US we have a shortage of general practitioner physicians, and we're trying to find ways to leverage AI and other next gen technologies to allow physicians to treat the whole patient in a rapid manner [and] where the patient sees complete and total care.
“And we get accelerated billing and, ultimately, reimbursement from the insurance companies because we use Gen AI to complete the billing.”
While Barton didn’t comment on healthcare billing in Australia, there are parallels – the complex interplay between Government insurer, Medicare, and private health funds was a tangle of billing for services delivery bogged by compliance and conflicting line items. Alone, the most recent Medicare Benefits Schedule (July 2024) — the Government’s ever-shifting list of subsidised health services — runs to 1589 pages, up 235 pages in five years.
‘Soup to nuts’: Gang of 100 to spur Unisys’ AI leadership
But careful and continuous integration to mesh people, processes and technology would be needed for maximum harmonious effect, Barton said.
While he singled out Microsoft Copilot as an “entree into a broadly based AI … for the masses”, speaking more generally about Gen AI he noted that “people struggle with prompt engineering; with writing a prompt that is complete or comprehensive enough for the generator to produce a response that is of value”.
Unisys stood-up a working group — the ‘Gang of 100’ — to get input from within. Barton said this has helped the veteran IT vendor, associated for many years with ‘Big Iron’ mainframes used by enterprises and governments, with its internal use of AI.
"This cannot be a technology-office-only operation,” Brett Barton, Unisys.
“We want to understand how [our] employees are operating. We are a technical-service organisation so our people are technically adept; why don't we just ask them how they'd like to leverage [and] interact” with AI?
“That has reaped incredible rewards for us as we become an AI-centric organisation, ‘soup to nuts’ [start to finish].”
And to derive the greatest benefit, organisations must shift to a programmatic mindset: “AI is like a child [that] requires nurturing care and feeding”.
“A lot of folks see it as a one and done; this is a program, not a project.”
Devil in the AI detail: Risks and rewards abound
As AI and associated technologies promised leaps in productivity and profitability, with great reward came great risk. Barton warned about organisations exposing themselves.
“[AI] is the first technology that can cause harm to an organisation if they're not careful,” Barton said.
“Clients have blocked ChatGPT, only to have employees take information on their phone and run it through just to see what comes back. And that information is now out in the wild. It's gone forever.”
As AI becomes more pervasive, Australian organisations must recalculate their “risk indexes”, Barton said.
“And for those companies that operate in regulated industries or environments, you are going to see more strict adherence to the risk models that are enforced by leadership of the organisation.”
But organisations that managed risk were in line to “create nuanced opportunity for … everything from employee health and safety to revenue recognition”, said Barton.
-1.jpg&w=100&c=1&s=0)
.png&w=100&c=1&s=0)
