Q1. Could you start by giving us a brief overview of your professional background, particularly focusing on your expertise in the industry?

I have been active in the Professional Shared Services industry for the past 29 years. I started as a project leader in an internal IT department of an American/Dutch manufacturer in the FMCG space. I became part of a comprehensive outsourcing and rebadge to HP, through which I joined the “provider side” of the industry. Within HP, I executed several roles, starting with large global project management, outsourcing delivery management, and account management. Having gained ample experience in the industry, I moved to Deloitte, taking a leading role in BPO outsourcing advisory in Switzerland, serving both Swiss and international organizations. I then continued my career with Genpact, first as Lead Client Partner and now, for the past five years, leading Global Advisory for all customers in the Manufacturing, Oil and Gas, and Chemicals industry. With my team, I help organizations build up their own Shared Services and GCCs, improve their processes, or automate the same with advanced technology and AI. My team and I do this for all enabling functions, including Finance, Procurement, Supply Chain, HR, and Commercial functions.

 

Q2. How are AI and automation reshaping the operating models of global capability centers and enterprise shared services organizations?

In my observation, I see that most traditional shared services organizations are somewhat trapped in the culture they´ve built up over the years in terms of standardization, compliance, structure, and focus on repetitive tasks. As a consequence, they are viewing Artificial Intelligence through a narrow, tactical lens: replacing human hours with agents to cut costs while keeping existing ways of working, the main processes, intact. Focus is on compliance, repetitiveness of the outcomes, and enrichment of the human being that essentially stays at the helm. Volumes of transactions are still driving the size of the centers, and AI has not yet been able to make a huge impact on the same. Stated differently, the impact that AI has been making on Shared Services so far has not been so impactful that the SSC would not be able to survive without it. We all see the potential value of AI; however, we all have a good grasp of the vision we try to achieve with it. Yet, our current ways of working stand partly in the way.

For Global Capability Centers, I see a different trend. With the focus of GCCs on more value-added tasks, highly contextual assignments, and innovation-focused work, the application of AI is different. Not focused on replacing human hours with agents, but providing new insights that no human being has been able to provide before, or for which a human being would have been collecting and analyzing data for months if not years. The AI has become an integral part of the value the GCC is delivering, a value it could never deliver at the same speed or with the same inclusion of contextual knowledge as the AI can. Meaning that without AI, the GCC would not be able to deliver on its mission and promise.

The above shows that there are two main use cases of AI currently: one where the human is being supported by AI, making tasks lighter and less effort-intensive. Work that is transactional in its core is being partly replaced by AI doing the work for you. And then there is the use of AI, which enhances human brain power and human analytics power. These are two different impacts on the operating model, and two impacts that should not be mistaken for one, as also explained in the next question.

 

Q3. How is the competitive landscape evolving between traditional outsourcing models, GCCs, and AI-powered managed services?

It has been a wish in the Shared Services industry and the BPO industry to become more business-relevant when it comes to their role in organizations. A relevance beyond cost improvement and efficiency. Business partnering. The materialization of that wish is, however, limited. SSCs are still seen as cost centers, who ´just´ need to execute on volumes of work being provided to them. Bring scale and efficiency to drive costs down. The GCC, on the other hand, is the new kid in town. The center where value is created, where investments flow into, all to drive true business value. Beyond enabling functions.

If I compare the positioning of the GCC industry versus the traditional outsourcing model, I see overlaps between the trend of about a decade back, between the Digital department and the traditional IT department. Digital required different thinking, a different culture compared to how IT went about its business. Digital was supposed to deliver value, being integrated in the company´s end-products to make an immediate impact on customers. And sometimes things broke: they failed fast. IT was supposed to manage costs and deliver stability. Don´t fail. Follow standards. Be auditable. Those are two organizations with two different missions. For that same reason, organizations created their Digital organizations alongside their IT department, not as a part of it.

I promote the same when it comes to the GCC. Meaning I´d be careful to try to turn a traditional Shared Service center into a GCC. Cultural clashes, different missions, and, at the end of the day, different styles and funding of doing business. GGCs should be built alongside an SSC.  Is the GCC with that a threat to the Shared Services? I don´t think so. It will “nibble away” at possibly a few elements, like expertise functions aimed at continuous improvement. Or maybe an AI CoE that has been built up inside the SSC. But it will not replace the Shared Services. Just like the Digital department never replaced the IT department.

That´s a different story for the AI-powered managed services. Within my customer base, I get more and more the question: “Should we invest in building up a shared services department, or should I invest in automation, and more particularly in AI?”. Today´s answer is still “Shared Services”, but the moment that answer will change is coming rapidly closer. At least, under the current commercial parameters of AI (more on that later in this mail). At some point, processes are running so much automated and autonomous that the added value of having an employee as part of a large shared service center is not that high anymore, and does not warrant the investment. That also means that if an organization already has a shared service solution in place, it will see its size shrink. Profiles of employees change from transactional champions to operators of multiple AI agents. Team leaders today manage a team of shared service employees, who are most of the time experts in how the process should be run. Those same experts run a team of mostly Agents who perform tasks that the main systems of a company don´t perform yet. I say yet, because Core ERP-system providers are integrating AI functionality as standard in their products. SAP recently announced the inclusion of about 40 Athropic agents in its products. Other cloud-based or “off-the-shelf” software solutions are on the same path. So bespoke, specifically designed. Agents to do a certain task on a legacy IT landscape will also be reduced in number. But they will never, at least not in the coming 10 years, be eliminated. Nor do employees.

 

Q4. How are companies balancing cost optimization with customer experience and operational agility in transformation initiatives?

Customer (and employee) experience and cost optimizations are not, by definition, two topics that need to be balanced. One can actually fuel the other. Especially when applying LEAN principles throughout the organization. In other words, building the transformation on the simple LEAN principle that waste is everywhere, is created everywhere, and needs to be eliminated. Doing so lowers costs significantly AND improves customer and employee experience. Companies with this LEAN principle at the forefront of their transformational initiative, and a rigorous application of the same throughout the organization, are not compromising one against the other. Look at the progression GE (and now its independent organizations) made in the past years after the arrival of Larry Culp. It´s an incredible turnaround that has been fully built on LEAN principles. Cost improvement, happy customers, happy employees, happy employee representatives, and even happy vendors.

When it comes to operational agility in combination with a transformation towards Shared Services and/or GCC´s, I´ve seen only positive stories. For me, Solvay is the use case that describes perfectly how an integrated and mature shared service organization actually empowered the better than modeled divestiture of one of its businesses. Or, less strategically, whenever new policies or ways of working need to be implemented, then having to address the same only once in your Shared Service center instead of having to address it to all your operating companies around the globe, massively increases your agility and your ability to change course quickly. The same, of course, being valid if the Shared Services are built on the LEAN principles.

 

Q5. What aspects of enterprise AI transformation do you believe are currently overhyped, and which opportunities remain underestimated, and what risks are not talked about sufficiently?

AI transformation: what is currently overhyped

With the IPOs of major AI players ahead of us, like OpenAI, SpaceX (with Grok), and Anthropic, we all see one very impactful use case followed by the next. This is my opinion to ensure that the public´s image of the provider and therefore its IPO stays extremely positive. And the use cases I personally see are indeed plausible. Maybe some after investments in a number of basics, and always some that are immediately working. Claude of Anthropic has built up a good, plausible stack of solutions that make life easier, and in some cases, are already partly replacing work that normally was being done by you and your colleagues. Think about the plug-in for PowerPoint presentations, for example, that automates parts of the workflow of producing a solid slide deck. However, what is currently not sufficiently underlined is that every model has its limits. And when those limits are reached, the model starts making shortcuts, or more commonly known as hallucinations. And real life is full of parameters that fuel the reach of those limits. The solution of limiting these parameters often comes back to standardization of processes or structuring of data. And here comes the catch: the more you need to standardize and structure to allow for more effective and predictable outcomes of the AI, the more “normal” automation comes back in sight. Meaning: due to the hype, we are tempted to throw AI at everything, but the reality is that many topics could have been solved with normal automation as well.   

The underestimated opportunities: Organizational memory

The market, in my opinion, is distracted by the hype of total human replacement and, to keep it more realistic, the enhancement of part of the workforce. What is not sufficiently talked about is the topic of using AI to build up organizational memory. We all have colleagues in our workplace who know how to request some home office goodies, who know who to call if you need help with a project you are doing, or who know the shortcut in the systems to keep the customer happy and to deal with the internals later. Those colleagues typically have long tenures. Some shorter. But altogether, we have become the memory, the instinct of the organization. Take one person out of the team, and the organization still works. Take 10% out, and the organization still works (commonly known amongst skeptics as “the McKinsey advice”). But take 20% out, and things start to fall apart. Here, AI can help by taking on that organizational memory. By allowing it to gain access to mail communication, TEAMS communications, documents, processes, ERP, and other system transactions, it becomes the organizational memory. And it allows for more drastic and less incremental restructurings.

The unspoken risks:

There are two risks associated with the use of AI that are not sufficiently talked about.

1. The real costs of using the AI. We all know that the current main AI providers like Google, Anthropic, SpaceX, and OpenAI are running at massive operational losses. They are funded by organizations that understand the opportunity and are willing to defy all accepted norms of valuation to be part of this gigantic global and human transformation. There will come a day when these investments need to start seeing a payoff. We already see the first indications of the same in tokenizing subscriptions or using pay-as-you-go models. My own personal Gemini Pro account just underwent such a change. And I bet everything on it that more is to come in that area. What I miss in today´s valuation of Business Cases for the use of AI is this exact price volatility. For example, by creating a business case with the “what-if” of massively increased costs for the use of AI, which would represent the real costs.

2. The all-in strategy. Let me take as an example the above topic of “organizational memory”. If an organization has gone all in on this strategy, choosing Anthropic as their platform, and training the platform at very high costs. What happens when Anthropic ceases to exist? That would be an unmanaged wipe out of all documented organizational memory. It would be like a massive restructuring in which 30% to 40% of the workforce is randomly let go. Operations stall, employees are lost without the help of the “colleague” who knew it all.

 

Q6. What signals indicate whether a GCC is evolving into a strategic innovation center versus remaining a transactional support organization?

By the definition of the GCC, at least how I defined it at the start of this mail, the GCC should have been set up on day one as the innovation center that is pivotal for the organization´s future. People with mission-critical knowledge supported by AI with mission-critical goals are, from day one, the “inhabitants” of the GCC. They are the capabilities the organization relies on to keep on developing itself. Not only to run it.

Earlier, I argued in so many words that rebranding a shared service center into a GCC is not the way to go. Both have their strategic importance: the SSC to keep the organization´s engine running efficiently, the GCC to be the provider of the new turbo, the new lubricant, the new hybrid engine variant, and the new cooling system of the engine. Or even the complete car around it. That´s the GCC versus the SSC. Both are important, both are pivotal for the company, yet with completely different missions.

Or to say it differently: for me, the signal that a SSC is on the wrong path is that it wants its parent organization to invest in becoming a GCC. 

 

Q7. If you were an investor looking at companies within the space, what critical question would you pose to their senior management?

In case I were to look at BPO companies, I would look at the following aspects:

To realize the ambition of becoming a Managed AI provider:

1. The Agent migration mix: without changing the underlying workings of a legacy contract, how many AI agents have you deployed (or AI-based solutions) that have replaced human effort, without having to allow for a reduction in contract revenues?
2. Proprietary IP: What proprietary platforms does your organization actually own, build, and scale, rather than simply white-labeling third-party tech?
3. Ecosystem Depth: What is the structural depth of your partnerships with the major AI platform providers? Is it a marketing alliance or a deep technical integration?
4. R&D Accountability: What exact percentage of your R&D budget is dedicated exclusively to AI engineering that would improve your current ways of working, and what specific capabilities is that capital building?

To realize the ambition of becoming a 3rd party GCC, I´d look at:

1. The existence of a GCC strategy that is not looking like a BPO on steroids
2. The organizational setup where GCC is recognizable as its own unique P&L
3. The investments in both human and AI capital for the benefit of creating GCC platforms where customers can tap into.

I would be careful with looking at parameters that act as derived parameters, like:

1. “Agentic revenues,” which is typically a mix of employees with modern (AI) technology. It is unclear what the actual income of the AI versus the FTE income is.
2. Non-FTE revenues, which can be a mix of own IP licenses income, third party license income, and, amongst others, fixed price contracts that are FTE contracts in reality.
3. Revenue per FTE (the higher the better), which is counterproductive in cases where the BPOs start offering more near-shore or on-shore personnel to improve this KPI.

These parameters are valid to keep in mind and valid to look at, but only address part of the story.

When I would be an investor in an organization that works on improving its competitive position by means of Shared Services and / or GCC:

1. Is the SSC setup indifferent to whether it is being in-house or outsourced, as long as the best mid-term results are being achieved?
2. Before investing in the SSC setup, did the organization do a study to invest the same dollars in further (AI-based) automation investments to reach the same outcomes?
3. Is there investment freed up for the SSC to become more of a managed AI solution, and if not, is the organization actively pursuing outsourcing the SSC to make the pivot towards a managed AI solution for them?
4. Is the setup of the GCC intended for innovation and provision of business / customer-facing solutions, or is it an SSC in disguise?
5. Does the GCC have sufficient funds to invest in cutting-edge technology to fulfill its mission of innovation?
6. What partner eco-system, especially a technology partner, does the company nurture to ensure it is able to continuously be a front-runner in innovation and technology?