Problem solving with AI at the Boston Consulting Group

As CIO of Boston Consulting Group, Merim Becirovic oversees the global technology strategy and enterprise transformation to solve problems within the company so it can, in turn, better solve those of others.And AI is integral to the entire conversation.

“I think every organization out there has to think about how AI is used internally and externally,” he says. “It’s a bit different for us because we’re a consultancy, so what we always want to do is prove things to ourselves first. In that sense, we become customer or client zero on many of these complex solutions in order to gain the confidence to figure out how to build things and show how we imagine a company running with AI should be.”

Becirovic recently kicked off a refresh of the company’s IT strategy for the next five years, encompassing a fundamental shift to position as an AI-powered company and learn along the way about what that looks like.

“When I talk about some core components, it’s really about how work is changing,” he says. “We’re starting to use these capabilities where we connect journeys across systems with AI tools, and once you start doing that, it starts to feel more consumer-like, versus just being yet another application or URL I can’t find.”

Deep in the throes of this IT journey, Becirovic, reflecting on his own career, is certain this is the best time ever to be in technology. “All those nights I tested data and scripts, uploaded code, and just did the work of machines, I think we’re at this space now where we’re finally giving machines the power to do the work of machines, and we as humans get to do what we’re great at, which is unlimited innovation, creativity, and all the different ways we think about solving problems.”

Becirovic also speaks about the importance of data strategy, upskilling, and managing AI costs effectively. Watch the full video below for more insights.

On enabling an AI workforce: Right now reminds me a lot of what we went through 10 years ago when cloud was the new technology. We didn’t have all the skills, but what was relevant back then that nobody saw coming was all the people in these spaces wanted to work on new stuff. They were investing themselves into building capabilities and growing skills. AI is no different, except it’s a lot faster because we’re just about at three years since the first ChatGPT was announced, and look at the amount of innovation since. We’ve moved from LLMs, to RAG models, and now we’re into agents. And then you start thinking about how these agents are going to disrupt everything, and then everyone’s going to have agents. So it’s changing very rapidly. I think the workforce itself is very excited about it. Most of my teams are. There’s no shortage of ideas in terms of what people want to work on. I think it’s really about trying to understand where the industry is shifting.

On data preparation: Data, of course, is the underlying prime factor if AI is going to work or not. Every organization has those challenges and we’re no different, but we’re very much focused on what our knowledge is, how and where we want to surface that knowledge, how we want to combine our knowledge with external knowledge, and how we bring all of those things together to help our clients.So it’s an effort to figure out what data products and platforms are needed, and how to start surfacing these things. If you have no data strategy, the biggest factor is getting the inventory and understand where all your sources are, and then start figuring out how to combine those things.

On skills: We’re going to start to see college students come to us with the agents they went through college with, like a personal assistant. This is coming. They’ll be asking how to use them at work. I think at the root or lower levels, you’re going to start to see that. But I’d argue even the definition of what a lower level is will change because many of these tools can now do a lot of rudimentary things very rapidly. College graduates are going to be much more sophisticated in having all these agentic capabilities as part of their portfolios as they come out. So I think that’s how you look at that junior level. At the senior levels, leaders have to understand and have to be enabled with these capabilities. And a lot of them already use them. We’ve leaned in, so we’re very passionate about that learning and capability. You have to open doors for people and give them the flexibility to try.

On cost control: I draw similarities when we went to cloud. Every organization that goes to cloud says they’re going to run it like a data center, but they can’t because costs keep going up. We had fixed costs and we’d buy a server, but in the cloud, that actually costs more money, so companies had to learn how to optimize and turn on schedules and work differently in the cloud to actually get the savings. And it’s the same thing with AI tools. It’s about deciding and understanding which models to make available and where, especially through applications. We have a lot of opportunity, and with some of the frameworks we’re building based on the type of questions being asked, we can go to a cheaper answer as opposed to a deeper research question for a client, where we’d want to expend an extra set of skills and not have somebody kick it off manually on their own. It’s important to get there, but what we’re all waiting for is all the ecosystem providers to integrate all our identity systems so it can be more manageable. If not, it becomes a free for all.