IBM Think: AI transformation aims to shrink process, lift people

AI transformation requires human-centric design

For legacy companies like EY, State Street, New York Life and Providence Health, the goal is to use AI to empower people by removing administrative toil, providing a better internal user experience and creating upskilling/career mobility opportunities.

Scott Berlin

“Any time anybody says ‘AI transformation,’ the natural inclination is to think of it as a technology project, but transformation at its core is really about people,” said Scott Berlin, senior vice president and head of group benefit solutions at New York Life. “We are going to redefine the interaction model between our people, our technology platforms and the new digital agents that we’re building. So, we are really going to rewrite every job description in the business over the next few years.”

New York Life has been working with IBM Consulting to transform its group insurance business. Berlin imagines a scenario where a digital agent could continuously review every file, detect when a physician refers a patient to another specialist and automatically request the necessary records without involving a claims manager. This frees up the claims manager to do more consultation and human to human interaction instead of paperwork in the background.

 “We’re really seeking to elevate our people though this process, into jobs that only people can do,” Berlin said.

At Providence Health, leadership shared a similar vision for AI’s ability to clean up process and let people do what they do best.

“We’re automating transactional pieces of work that we want you to stop doing,” said Carol McDaniel, vice president of talent acquisition at Providence Health. “We want you to do more intense patient care. We want you to look after our caregivers.”

Carol McDaniel

Providence Health has been a longtime partner with IBM through its business process outsourcing function, where IBM oversees the payroll function, talent acquisition and HR service. Providence worked with IBM Consulting to deploy their AI-powered requisition and internal transfer agent (Rita) using IBM watsonx Orchestrate integrated with its existing HR platform.

“Specifically talent acquisition is ripe for automation and ripe for agentic AI,” McDaniel said. Providence Health used Rita to make the internal talent mobility processes faster and more curated to individual roles. The human-centered design approach involved designing the process and experience first, then mapping the technology onto it afterwards.

“I would not dare introduce any type of agentic process without having went through that process,” McDaniel said. “Our leaders can actually transfer people through the process 12 days faster than they were doing before. And the candidate or carrier is still experiencing that valued, seen-and-heard experience we want them to.”

How roles might change

At these regulated companies, AI is removing some work while increasing the demand for new skills.

“Sure, [AI] can replace the person who’s pushing buttons all day and having to do things. But at the end of the day that person also brings value at a different level to the organization, and if you’re not investing in that talent, then it’s a miss for your organization,” McDaniel said.

Organizations are rethinking where employees add the most value in increasingly AI-driven workflows. Rather than eliminating the need for people altogether, executives said AI is elevating the importance of strategic thinking and judgement.

Christopher Aiken

“We are actually experiencing the need for even more human intellect on several levels,” said Christopher Aiken, principal, tax technology services at EY. “One is, quite frankly, just a stronger focus on planning and decision-making that today is perhaps in short supply because of all the time that is spent on manipulating data and cleansing information.”

Aiken also mentioned how AI is creating new opportunities in the tax business, citing the replacement of payroll tax for humans with taxes on AI and digital workers.

“That is a new opportunity for us as a business, and we need our new workforce to be thinking about this opportunity,” Aiken said.

New York Life uses digital agents in the background, not customer-facing, though Berlin says customer-facing agents aren’t out of the question.

“I imagine every employee sitting in a command center, where insights are being surfaced to them at just the right time,” Berlin said.

I imagine every employee sitting in a command center, where insights are being surfaced to them at just the right time.

The thought is that, by having work automatically surfaced to you, you decrease the cognitive load that comes with searching for, keeping track of and maintaining information.

“AI is replacing the transactional or administrative work that we get tied up doing, that I even have to do myself, and releasing people from that burden,” McDaniel said. “[This gives] them the opportunity to be a strategic partner in the business and to think a little bit more about designing a better work process.”

McDaniel cited AI as a potential solution to the hiring shortage in healthcare, by upskilling more entry level roles like diagnostic imaging, CNAs and medical assistants. “If you can take the million plus dollars you’re saving by removing the transactional work, and reinvest those dollars in upskilling your workforce, you just trained your internal workforce, your own internal hiring opportunities.”

Tech enablement and upskilling

In addition to upskilling or providing opportunities for talent mobility, leaders want to encourage employees to adopt AI and for it to become a natural part of work.

New York Life lets all its employees access ChatGPT. The company has a group transformation office that helps employees find ways to use AI in their own work and connect other employees using it.

“I don’t think using ChatGPT should feel any different than using a stapler,” Berlin said. "None of us fold the ends of the staples by hand, we use a machine for that. Now if you’re writing a memo, why not have ChatGPT there to help you iterate and build it better?”

Aiken explained that the average median age of a tax director at a Fortune 500 company is dropping significantly, and that their expectations around tech are higher.

The user experience our workforce is expecting is one of extreme tech enablement.

“The user experience our workforce is expecting is one of extreme tech enablement,” Aiken said. “We hear that on campus, and we do a lot of advocating for that on campus. We are bringing those technologies to our users.”

That expectation is also present on the client side.

“Our clients also expect a high-tech experience from us -- even doing something as traditional as accounting and tax,” Aiken said.

EY’s business structure was actually a plus for AI adoption.

“We’re very lucky in that our entire business was rooted in an apprenticeship model,” Aiken said. “That model actually works very well in a mixed human and AI environment, because it allows us to introduce AI in a way that leads to the proper level of review [and] human oversight. It’s sort of a built-in safety mechanism of how our business operates.”

Challenges of AI transformation and a look ahead

Despite the purported benefits, executives acknowledge the significant change necessary to adopt agentic AI across the organization, especially in legacy organizations such as EY, Providence Health and State Street. These organizations have spent decades codifying certain processes and ways of thinking and employ people from every generation.

Communication and experience

Executives expressed the importance of consistent communication to remind workers – and themselves at times – of the AI-first vision.

“This is a multi-year journey we’re on. It’s not like we’re going to build it in May of ‘26 and launch it next month,” Berlin said. “Any time you’re on a multi-year journey, people will have anxiety about [what it means for them].” “Open communication, transparency and showing people the vision of where we’re going is critically important. Communicate often -- not just once -- but throughout the process.”

Cleaning up processes and data

McDaniel explained that a lot of background work went into creating Providence Health’s internal transfer module, where users can click a couple of buttons and the AI tool Rita helps them move through the process. This involved cleaning up stores of old job data with thousands of open postings that hadn’t been filled in a decade.

“We can’t just automate a bad process. We have to go back, clean up and redesign the process, which includes cleaning up the data,” McDaniel said.

We can’t just automate a bad process. We have to go back, clean up and redesign the process, which includes cleaning up the data.

This ground-up process revision is what sets companies apart in the AI race.

“The next phase of AI will not be won by companies with the most AI tools or copilots. It will be won by companies that redesign how work gets done,” said Vuk Janosevic, a Gartner analyst. “AI-native does not mean sprinkling AI across existing workflows. It means rebuilding the operating model so humans and AI can jointly plan, execute, govern and improve work.”

Where AI does not belong

Some tasks just aren't fit for AI, no matter how AI-native the organization is.

“I am not thinking about AI making decisions that people should make,” Berlin said. “If it’s a simple approval, I’m comfortable having an AI approve it. If a woman files a maternity claim with us, we’re going to approve that 99.9% of the time, unless it’s something strange. If we have a digital agent doing that approval, I think we can do that and [it’s] probably a better experience.”

Even the most AI-forward companies recognize there needs to be a place for humans. “[Humans] will always be in play at State Street because we’re extremely regulated, A, but B, we know that at the end of the day, human-in-the-loop will bring the soft touch to the completion of any work that we do,” said Manoj Bohra, chief data & AI officer at State Street.

SaaS remains but will evolve

Just as AI will not replace people, it will also not replace the tools people have depended on.

“I don’t think we’re close to replacing the core systems. I do think the interface will become less important over time,” Berlin said. “I don’t know if in the future our distribution folks will be typing things into Salesforce. I think they’ll be talking to their digital agent, which makes that interface not valuable at all. It doesn’t need to be there.”

Bohra shared a vision where the interface is geared toward agentic communication instead.

Manoj Bohra

“In the new economy, what will happen is that the agent-first transactions will overtake,” Bohra said. “It is critical that we set up a foundation for State Street where we can reach a ... future where agent-to-agent communications, transactions, execution of workloads can happen.”

However, that doesn’t detract from core platforms’ importance. At New York Life, doctors and clients both use tools like Salesforce and SAP to perform their jobs, store documents and send messages.

“We don’t need the same interface with Salesforce, but that doesn’t mean we don’t need Salesforce,” Berlin said. “I think that these platform companies will have to reinvent themselves a little bit [to preserve the important part of what they do] five years from now.”

Berlin explained that their AI layer is built on top of the existing claims or underwriting systems.

“The future of enterprise AI depends less on better models and more on better orchestration: the ability to coordinate people, systems, data and AI agents under clear rules,” Janosevic said.

AI needs to understand the data in the context of the business, and that data, with context, partially resides in traditional SaaS tools.

“Traditional software helped people do the work. AI-native systems increasingly help execute the work,” Janosevic said.

Although the future of enterprise AI is uncertain, leaders are driven to invest.

“At some point the intersection of reality will happen,” Berlin said. “But we really want to think big and bold in this moment.”

Ben Lutkevich is an award winning technology writer and editor covering IT infrastructure, app development and AI.