Most CIOs deploy technology created by vendors or developed internally for their own organizations. At Boston Consulting Group, CIO Merim Becirovic operates in a different environment. Every day, the firm’s consultants help clients across industries rethink business processes, redesign operating models, and explore new applications of AI. That gives him and his team visibility into a constant stream of experimentation and innovation happening beyond BCG’s walls.
Rather than viewing those ideas as client work alone, Becirovic embraces the concept of customer zero, applying the most promising capabilities within BCG itself. The result is a powerful feedback loop: client engagements generate new ideas, BCG tests and refines them internally, and the lessons learned strengthen future client work.
In this conversation, Becirovic discusses how BCG scales AI through its Genie platform, gives employees the freedom to experiment safely, and turns emerging ideas into the transformation of how work gets done.
What impact is AI having on businesses?
AI brings an upskill of capability and opportunity to do more with technology than ever before. It’s a fundamental reimagining of how tech enables business. For years, we built systems based on the constraints of technology, because we couldn’t get what we wanted out of performance, storage, data, or networks. IT was expensive so we focused on efficiency.
Then we got off the mainframe and into distributed computing, big data centers, virtualization, and the cloud. While AI has existed for years, we finally built the capacity to apply it.
Until now, technology practitioners have done the technical work because it was too costly to get machines to do it. With AI, they can now do the work of machines, meaning you can free humans to do what we’re good at: creative thinking and using AI as a capability to enable that.
How is BCG applying AI to its own operations?
We use the concept of customer zero, and apply the capabilities we imagine for our clients to ourselves. I’m lucky because I have access to consultants around the world who think hard about how to help our clients answer the same questions we’re asking: How do we become an AI-first company? How should we think about deploying agents in our own organization?
Maybe we want to look at contracting differently or use AI to power our engineering teams. That might be something we do for a client, but I want to promote customer zero so we gain that knowledge and apply it to us. This is how we get better using these foundational models.
When work is changing and new technology is available, there are many more iterations of ideas that might be interesting, and they happen a lot faster. Everybody’s been experimenting with custom GPTs and agents, but only as standalone capabilities. As an enterprise leader, I want to take use cases out of project mode and bring them into enterprise scale without creating technical debt.
What’s an example of a scaled AI capability at BCG?
BCG Genie, which won a 2026 CIO 100 Award, is the platform we use to take AI ideas out of project mode and run them across the enterprise. It gives us one secure place to build, govern, and scale agents and other AI tools, using what we developed for clients through BCG X, the part of BCG that delivers technology and data capabilities.
Today, everything we do to support BCG, from HR to tech support, runs through Genie, with more than 100 use cases. Key to Genie is Spark Lab. You have a spark? Here’s a lab. If you’re a consultant who wants to test whether something’s a good idea, you put it in a sandbox, connect it to additional information in the business, and start to get insights. Then you think about how to roll an idea out to the whole company if it’s really good. Spark Lab provides a self-service idea generation and scaling opportunity for everyone.
Some CIOs will fear people trying technologies that aren’t part of their tech stack. But Spark Lab users don’t use new technologies. They’re consuming patterns from the platform, and we apply security and data controls as the idea scales. We don’t refactor up to our standards later; the standards are there from the start.
What do CIOs need to get right to make AI pay off?
Technically, you need a solid and well running standing in the cloud. But on the people side, it’s giving people agency to use the tools, not just give access to them. We’ve solved so many of the hard problems of technology, our new job is to change the way work is done.
My peers know we’re change leaders now, not just technology leaders. Our vantage point across the business lets us connect the dots and drive transformation.
In a real transformation, 30% is the models and the technology, and 70% is people and processes. Most organizations stop at the 30%. The question is how the 30% affects the 70%, and how you transform and invest in that 70%. That’s how you get compounded benefits.
What’s the difference between alignment and agreement?
You can get alignment, but do you have agreement? We can all align on being AI-first, but agreement means we’ll build these platforms, use these capabilities, transform these workflows, and change the way we work. Without agreement, you get a lot of false starts. It’s a subtle difference, but with AI, it matters more because the work itself is changing. We have to agree on how we take advantage of these capabilities.
What’s your advice for emerging CIOs?
Be hyper curious. Play with all the technology and look at different ways you can bring it together to solve business problems. Even with AI, enterprise technology is hard. It won’t magically connect everything, so understand the trade-offs and complications. We all still need to do the brilliant basics of running an enterprise that serves many different types of work.
The more experience you can get around how technology is delivered, used, reimagined, and applied, and how we evolved to AI, the better set up you’ll be as a next-gen leader, because this stuff is moving so fast.
And remember the 70%. Many companies have paid a lot for half-completed transformations. The technology mostly worked, but something happened in the adoption cycle. Technology will become easier and cheaper to consume. We can make mistakes faster and more cheaply, and refactor code and capabilities. But changing how people work is still an evolving capability, and we must keep investing in it.