From performance reviews to pink slips, managing AI agents looks a lot like managing people

The workforce is no longer purely human — and closing the gap between how companies manage people and how they govern AI agents has become one of the defining operational challenges of the digital worker lifecycle.

As agentic AI spreads across every facet of the enterprise, IBM Corp. is betting that the same management rigor applied to human employees must now extend to digital workers. The core of that thesis is the digital worker lifecycle — a structured approach to hiring, credentialing, deploying and retiring AI agents with the same discipline organizations that has long governed people management, according to Mohamad Ali (pictured), senior vice president and head of IBM Consulting at IBM. That perspective started when IBM CEO Arvind Krishna charged Ali with building a software layer capable of managing human and digital workers side by side.

“Arvind called me and we had a conversation about three years ago and he said, ‘I want you to come over and I want you to be a consulting business,'” Ali said. “‘What is digital labor? It’s a whole bunch of bits of software. I need you to come in and I need you to build this set of software and do it in a way that it could be managed. You could think of this as HR management … but now you have to HR manage human workers and digital workers.'”

Ali spoke with theCUBE’s John Furrier and Dave Vellante at the Think 2026 event, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed the digital worker lifecycle, AI-driven business transformation across IBM’s client base and how IBM Consulting is applying its own client zero methodology to prove out enterprise AI ROI. (* Disclosure below.)

The digital worker lifecycle — from hiring to retiring

The IBM framework moves well beyond basic agent deployment. The IBM Consulting Advantage platform now runs more than 4,000 digital workers across 450 active projects, operating on infrastructure that includes AWS GovCloud and other secure environments. The approach gives individual teams freedom to build on any AI stack — IBM watsonx, Anthropic, OpenAI — while routing everything through a common management layer that provides full observability and control over the agent fleet, according to Ali. The agents who then go unused don’t linger — they get cut off.

“If you build an agent that nobody’s using, eventually we’re going to decommission it,” Ali said. “We’re going to starve it. It’s not going to get tokens, it’s going to retire.”

The credentialing dimension of the digital worker lifecycle is where IBM’s work with Pearson PLC breaks new ground. Rather than testing agents on memorizable material, IBM and Pearson are applying workflow-centric assessments through Pearson’s Credly platform to issue skill badges — such as cloud essentials or security — directly to AI agents, Ali explained. The approach was developed with Dave Treat, who leads the credentialing initiative at Pearson, and solves a core verification problem.

“You can’t just give the agent the textbook — it’ll just memorize it and get all the answers right,” he said. “What Dave is doing is a much more sophisticated way — giving the agent problems, workflow problems it’s never seen before. He doesn’t have to grade it on multiple choice because he’s got an agent grading an agent. You can actually have our agent do a complex workflow thing and another agent grade how it did that in that workflow — a workflow that it’s never seen before.”

The business case for getting this right is concrete. IBM’s own consulting business expanded profits by 20% from 2024 to 2025 after decomposing its operations into 490 workflows, re-engineering 70 of them and applying AI systematically — a playbook that is now showing up in client results, Ali noted. U.S. non-profit Providence Health & Services, which deployed watsonx-powered HR agents integrated with its existing Oracle Corp. infrastructure, is now recruiting nurses 12 days faster, a sign that workflow-level AI transformation — not broad license rollouts — is where the value actually lives.

“We took a $25 billion spend and we’ve actually [saved in productivity] four and a half billion of that spend,” Ali said. “That only happened because we decomposed our company into these 490 workflows, took 70 of them, re-engineered them and did it the hard way.”

Stay tuned for the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of the Think 2026 event:

(* Disclosure: TheCUBE is a paid media partner for the IBM Think event. Neither IBM, the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE