AI + HI = The Only Equation That Matters in Shared Services 

A note before you read this.   
This article is not a research report. It is what happens when someone with nearly two decades in and around shared services spends six months listening to the brightest minds working at the intersection of AI and workforce transformation, testing their frameworks against real work, and trying to put the most important pieces together in one place. I have read the research, attended the conferences, listened to the podcasts, and sat in the practitioner sessions. What follows is what I made of all of it. I hope it saves someone else a little time on the journey. 

I have been working in and around shared services since 2006. For nearly two decades, I have watched  the model absorb wave after wave of transformation pressure, ERP consolidations, offshoring decisions, robotic process automation, digital overhauls, and now artificial intelligence (AI). Each wave brought the same promise and a familiar pattern. The technology arrived. The business case was built. The rollout happened. And somewhere between the launch announcement and the annual review, the results landed somewhere between disappointing and hard to explain. 

AI is different. Not because the technology is more powerful, though it is. Because this time, the barrier was never the technology. 

Here is the question worth sitting with before your next AI initiative gets approved. It is 2026 outside of work. What year is it inside your organization? Jason Averbook asked that question, and it is one of the most clarifying provocations I have heard in years. Forrester puts consumer AI adoption at 62%. Worker adoption inside organizations sits at 26%. That gap is not a technology problem. It is not a training problem. It is something deeper, and understanding what it actually is makes all the difference between organizations that transform and organizations that just spend. 

My vantage point is HR Shared Services, where I have led and lived this work. But the equation I want to share is the same whether your shared services function handles people, finance, procurement, or all of the above.

AI + HI = The Only Equation That Matters. 

AI is the capability. Human Intelligence (HI) is what makes it mean something. And in shared services, where the work is high volume, process-intensive, and deeply connected to the humans it serves, you cannot have one without the other and call it transformation. 

Adoption Is Table Stakes. Maturity Is the Game.

Most shared services organizations are measuring the wrong thing right now. License activation rates. Training completion percentages. Usage numbers. Those metrics have a role, but they are not the destination. They are the starting line dressed up as the finish line. 

Jason Averbook describes three stages every person and organization moves through with AI. Using it, where the tool exists in your world but not yet in your workflow. Adopting it, where habits are forming, and the tool is changing how you actually work. And Embodiment, where you cannot imagine working the way you used to. The context switch is gone. AI and work are the same thing. I know this because it happened to me. At a certain point, the question stopped being whether to use AI and started being how I ever worked without it. 

Wipro and SSON's 2025 CXO research found that only 13% of organizations have deeply embedded AI, and only 18% have achieved enterprise-wide redesign. Those are not laggard numbers. That is the field. Gallup's Q1 2026 workforce study of more than 23,000 employees found that while 50% of workers now use AI and nearly two-thirds report individual productivity gains, only one in ten strongly agree that AI has transformed how work gets done organizationally. The gains are staying at the task level. They are not reaching the organizational level. 

The winner in this era will not be defined by adoption. They will be defined by maturity. 

For shared services leaders, this is both the challenge and the opportunity. Your function was built for process discipline and measurement rigor. You know how to run toward an outcome. The question is whether you are measuring the right one.

The Obstacle Was Never the Technology. 

Boston Consulting Group introduced a concept worth knowing: the Silicon Ceiling. The organizational barriers that prevent employees from applying AI capabilities they already possess. Governance policies. Approval processes. Cultural norms around experimentation. Risk frameworks that treat curiosity as a liability. These are the barriers. And you cannot train your way past a structural problem. 

Ethan Mollick of the Wharton School adds another dimension. The tools can do more than most people realize, but most organizations are making everyone access them through chat windows, which are terrible for complex work. The people who struggle most are the ones who need help most. Interface matters as much as the model itself. The best AI in the world does not deliver value if it cannot connect to the systems people already use or if the integration layer is badly built. 

This reframes the transformation conversation entirely. Shared services functions have spent years becoming experts at process design, service delivery, and technology integration. Those capabilities are not less relevant in an AI world. They are more relevant. The question is whether the function is applying them to its own workforce transformation with the same discipline it applies to service delivery.

The Human Side Has to Be in the Room. 

Here is what I keep seeing across the research and the practitioner community. Organizations design the AI solution first and bring the people side in afterward to manage the transition. That sequencing is the mistake. 

Ecolab, ranked ninth on Fortune's AIQ50 and one of SSON's Top 20 Most Admired GBS organizations globally, redesigned their entire lead-to-cash process using nine agentic AI agents across a two and a half year journey. When asked what they got right, their answer came down to three things: lead with enterprise value, reinvent the process from first principles before you automate it, and if you do not rewire behaviors and ways of working, you will not find the value. That is not a technology story. That is a leadership story. 

Research from Teamraderie, in partnership with Stanford's Bob Sutton and UC Santa Barbara's Paul Leonardi, puts precise language to why this matters. Substitution, doing tasks faster with AI, produces a productivity gain. Recomposition, rethinking how work is divided between humans and AI toward a shared outcome, is transformation. Approximately 80 to 95% of organizations investing in AI see no measurable enterprise-level ROI because they stop at substitution and never reach recomposition. Individual tool usage is up almost everywhere. Workflow change at the team level is not. 

Shared services functions are particularly exposed to this risk. When your entire model is built on efficiency and throughput, the temptation is to deploy AI as a faster version of what already exists. The harder and more important question is whether the process should exist at all in its current form. What would this work look like if we designed it today, with AI native to the design, starting from a blank page?

What Operationalizing the Equation Actually Looks Like. 

AI + HI is not a philosophy. It is a design requirement. And in my experience, the shared services functions getting this right are the ones treating workforce transformation as a structured engagement, not a change management afterthought. 

The engagement I have seen work moves through four phases that are deliberately iterative. 

The first is Discover. Nothing gets designed until you understand what is actually changing. That means assessing the use case against business strategy, not technology capability. It means establishing a behavioral readiness baseline across the affected team before the change begins. And it means mapping the workforce impact with specificity, which roles are being redesigned, which are being augmented, and which early career pathways are at risk. 

The second is Design. This is where shared services have a structural advantage. Process discipline, measurement culture, and service design expertise are exactly the capabilities that rigorous workforce design requires. Org structure, job architecture, change leadership plan, and a blended workforce model all need to be co-built with the business, not delivered to it after the technology decisions have already been made. 

The third is Deploy, and it is where most organizations declare victory too early. Plans meet people in Deploy. The work is not done when the plan is launched. It is done when behaviors are changing, the structure is holding, and every affected employee has clarity about where they are going. 

The fourth is Evolve. AI use cases do not stand still. The shared services function that deploys and declares success will find itself re-managing the same transition eighteen months later when the capability has moved again. Evolve is where the engagement becomes truly iterative, re-measuring behavioral movement against the original baseline, updating the workforce structure as the use case matures, and keeping the long-view workforce decisions visible and active rather than decided by default. 

Running through every phase, four persistent decision conversations that must never be bypassed: what does the blended workforce actually look like, who gets reskilled or redeployed or reduced, how does job architecture need to change, and what is the deliberate plan for the early talent pipeline if entry-level roles are automating. These decisions do not get answered upfront. But they have to stay on the table throughout.

The Function Responsible for Humans Cannot Sit in a Supporting Role. 

Shared services, and HR Shared Services in particular, sit at a unique intersection in this moment. The biggest barrier to AI transformation, as the Wipro and SSON research confirms, is not distrust of the technology. Only 8% of leaders cite that. The real barriers are replacement anxiety at 37% and workflows not designed for human-AI collaboration at 31%. Those are human problems. They are cultural problems. And they belong to the function responsible for humans. 

That does not mean telling the business what to do. It means being in the room when the work is being redesigned, not called in afterward to communicate the change. It means applying the same rigor to workforce transformation that shared services applies to service delivery. And it means holding the line on the human side of the equation, even when the pressure is to move faster than the people are ready to go. 

Microsoft's 2025 Future of Work Report introduced a term worth sitting with: workslop. AI-generated content that appears useful but lacks substance. Forty percent of employees received workslop in the past month. Fifteen percent of all work content is estimated to be workslop. This is what happens when you measure usage instead of value. The quality layer, human judgment, critical thinking, and the ability to recognize when AI is confidently wrong, have to be deliberately designed in. The tool does not know the difference between good and good enough. That is still on us.

The Finish Line.

Nearly twenty years in shared services has taught me that the functions that endure are the ones that earn the mandate rather than claim it. AI is the most significant test of that principle in the history of this profession. 

The technology is not the answer. The technology is the condition. What you do with it, how you design the human side of it, how you build the culture and the structure and the judgment to make it mean something, that is the answer. 

AI + HI is not a formula for your next business case. It is the operating philosophy for the next decade of shared services leadership. The organizations that internalize both sides of that equation and build the discipline to hold them together will be the ones still standing when the next wave arrives. 

The winner will not be defined by adoption. They will be defined by maturity. 

And maturity is not a technology decision. It is a human one. 
 

Tom Sterling is an independent thinker on AI, leadership, and the future of work, with nearly two decades of experience in and around shared services. He writes and speaks on AI workforce transformation, the human side of organizational change, and what it takes to lead through the shift.