How IBM Transformed Its HR Function with AI

It was a 2017 staff meeting, IBM CHRO Nickle LaMoreaux recalls, when her predecessor first proposed having a digital chatbot rather than HR business partners answer routine inquiries from front-line managers. “There were maybe 10 of us in the room, and I was the only person who voted against it,” she remembers.

Yet despite that initial skepticism, eight years later LaMoreaux is now one of HR's most influential advocates for the possibilities of AI. She spends at least 25% of her time speaking with CHROs, in public forums and with IBM clients, about AI’s potential. The publication HR Executive credited her with “modeling innovation for the entire enterprise” on AI when it gave her its leader of the year award in 2024. And at the SXSW event in March, a line formed out the door for her panel about AI agents.

The AI-driven HR transformation begun by her predecessor—which LaMoreaux expanded and now leads—has had a major impact on the size and shape of certain roles on the team. Over a seven-year period, the company slowly went from having 1,500 HR business partners serving 29,000 managers and executives to just 350 HR business partners serving roughly the same size group. This was largely done by redeploying HRBPs to take on more complex work, not backfilling voluntary attrition (LaMoreaux says she runs a 10% attrition rate through retirements and voluntary departures), and some long-term performance management.

While the AI chatbot faced some initial resistance, it now boasts high recommendation scores, more than 11 million interactions a year, and a 94% “containment rate” of transactional queries about things like sick time, HR policies, or benefits questions that shouldn’t have to be routed to professionals. (LaMoreaux is careful to note that sensitive coaching or management issues like low performance or misconduct are directly routed to HR subject matter experts.) “We didn’t do this perfectly,” she says. “But if I can help other companies get there or make sure they don't make the same mistakes we did, I’m all for that.”

Modern’s editorial director, Jena McGregor, spoke with LaMoreaux earlier this year about what she’s learned. Here are 10 takeaways from the conversation:

1. Crises spark fundamental change.

In 2017, the IBM HR function was tasked with cutting its budget by 25%, says LaMoreaux, after a decade of already cutting 5 to 10% a year. “If there was low-hanging fruit, it was gone already,” she says.

At the same time, two other factors were competing for that budget: Local laws were becoming more varied, adding to greater HR management complexity, and employees were using more consumer-friendly apps in their personal lives, bringing those expectations to work. “Both of those things cost money we didn't have,” LaMoreaux says. “We had to do something fundamentally different.”

2. Sometimes, behavior change requires force.

IBM already had an early version of a chatbot in place, but after a year, no one was using it. “Why would they when they could walk down the hall and ask their HR business partner the same question?” LaMoreaux asks.

They decided to go “big bang,” as LaMoreaux refers to it. Overnight, the team took away the email and phone number for HR, and for 21,000 first-line managers, they also removed their direct access to HR business partners. From that point forward, the AI assistant was the only way to reach them.

3. But be patient for adoption—and quickly incorporate feedback.

The initial response wasn’t positive. Net promoter scores, which measure a user’s likelihood to recommend, plummeted from +19 to -35 on the initial rollout. Once they incorporated employee feedback, and the AI started getting “smarter” as managers and employees used the tool for 18 months, NPS scores bumped up to a +35, and currently sit at +74, LaMoreaux says.

After about two years of use, she noticed a tipping point. At first, “you're fighting the resistance, the fear of change. After that 24-month mark, I couldn’t stop the ideas coming in to my team. But you do have to get through that first rough road, whatever that period is in your organization, where you feel like you're constantly selling it.”

4. Start early. Time can help avoid mass actions…

Initially, HRBPs were all reassigned to help with more specialized questions. Over time, LaMoreaux says, she avoided large-scale cuts by relying largely on not backfilling natural attrition, saving budget by outsourcing more contracts, managing out some low performers, and having some staffers self-select out of the more complicated work.

“Getting started early allows you to be transparent about where the transformation is going, and then people have the agency to make their own choices.” She likes to remind people “we are seven years into our journey. That didn't happen in one night.”

"Getting started early allows you to be transparent about where the transformation is going, and then people have the agency to make their own choices."—IBM CHRO Nickle LaMoreaux

5. …while strong data buys you confidence.

IBM had done a time analysis study which found that 65% of the work HR business partners who supported front-line managers did was transactional, answering questions like “what's the maternity leave policy?” That gave them confidence a digital assistant could make a real impact. Plus, they could show that HRBPs were spending too much time on routine work—and doing it for too many employees, leading to lower engagement scores for HRBPs and unenthusiastic net promoter scores from the managers they served.

Later, when LaMoreaux expanded the effort by asking another 8,000 higher-level managers and executives to also use the digital assistant, she reminded executives that more than 90% of them were already using it, even though they still had access to HRBPs. Armed with data, “I was able to say, ‘this move isn't as dramatic as you might think.’”

6. Talk to your executive team early and often—and be clear on objectives.

When making a pivot in HR service with the executive leadership team, LaMoreaux says, she was careful not to surprise them. “I did about seven months of constant conversation with the ELT. It gets a little more personal.” It’s also critical to be clear on objectives. Leaders said to her: “ ‘Nickle, you have a hundred other line items in your budget, go take from other areas that don't affect me,’ ” she recalls. Telling them “I can give you better, faster service” too really helped.

7.  Look for the “legends.”

Start with the questions that are already flooding your HRBPs’ inboxes, she says. Indeed, LaMoreaux says if she were to start over again, she’d first use the chatbot for high-volume, simpler issues, rather than the full-scale rollout they took.

Then, to really drive adoption, find the things people complain about most or the “moments that matter,” she says. For IBM, one of those was generating employment verification letters, which can cause big headaches if there’s a mistake, but also often get requested at the last minute, such as when an employee’s child wants one for a FAFSA application due at midnight. “The employment verification letter became like legend at IBM,” she says. “It became this moment where people were like, ‘oh, did you know it could do that?’”

"When rolling out AI tools, remember it’s the process, the service delivery model, the culture. If you can’t do those other things, it will never work.”—LaMoreaux

8. Anticipate the objections.

LaMoreaux says CHROs typically see three kinds of objections. The first is philosophical: “First-line managers are the company,” people might say. “They are the front line, and we are abandoning them?” The second is logical. She herself was an early naysayer, and as the head of total rewards when the effort began, LaMoreaux worried that if HRBPs were taken away from first-line managers, “my compensation team was just going to get inundated with questions. I didn't believe we would re-engineer the workflows.”

The third is personal, “or what I’ll even call egotistical. I was offended. HR is my profession, and you’re telling me it can be done by a chatbot?” Recognize these are common and be prepared with responses.

9. AI agents are the future.

LaMoreaux says for IBM, her priority now is having a single AI experience and workflow with AI agents, so that whether you’re working on recruiting or transferring an employee or updating an address, it’s one experience.

Agents, which are systems or programs that can autonomously perform tasks on behalf of a user or other system, help integrate that experience as they take on more complicated tasks. “If I say I'd like to initiate the promotion cycle for an employee, it will start walking me through that [if] I'm authorized,” she says.

10. Finally, don’t call it just a digital transformation.

As you’re implementing AI in HR or using it to reimagine the work or size of your HR team, don’t call it a tech transformation, LaMoreaux warns. “It’s the process, the service delivery model, the culture. If you can’t do those other things, it will never work.”