Telcos face a cost dilemma: ballooning capital needs and commoditized pricing pressure. Standard cost reduction levers like simplifying individual processes or making procurement smarter are reaching their limit at many organizations. Enter AI. Yet while the opportunities from AI are highly promising, few telcos have realized value at scale from the technology. Marcus Wittig, a BCG managing director and partner who works extensively with telcos, explains how leadership teams can think through these issues, implement AI with a value creation mindset, and make the cost improvements they need to remain competitive.
Meet Marcus
BCG: What are the biggest cost challenges that telcos face right now?
Marcus Wittig: One of the key issues is many don’t apply an end-to-end focus. Telcos often default to levers that they can execute within a single function, like energy efficiency, capex optimization, customer service, and procurement. These work well—to a point—but they’re often not enough on their own anymore. Telcos and their cost structures are inherently cross-functional. Without coordinated effort across departments, companies struggle to have a real impact on costs.
Can you give an example of how this approach falls short?
Take a process like B2B order-to-cash. It spans sales, operations, and finance. Companies might try to cut costs in service alone. But if sales and finance aren’t aligned, that won’t lead to real impact, and the customer experience is going to suffer.
Or when a telco tries to streamline its service offerings to reduce complexity, a low-volume product may appear ripe for elimination especially if it has high service and delivery costs. But sales teams often push back, citing that product’s importance to strategic accounts and continuing to sell it.
In both cases, costs won’t come down—or worse, will reappear in a downstream function.
How can AI help telcos solve those problems?
When it’s used well, AI is a game changer in how to manage cost. (Of course, it also creates opportunities to add revenue.) AI provides opportunities within functions: customer service is a great example where many telcos already use AI to improve their productivity and realize cost savings. In addition, AI helps dissect and address inefficiencies across the value chain, from marketing to procurement to service.
When it’s used well, AI is a game changer in how to manage cost.
In functions like B2B, where telco offerings, products, and service delivery are complex, AI, in particular agentic AI, helps manage the many moving parts. On the front end, telcos can use AI to streamline how customers order, pay, and request support. On the back end, AI agents can reconcile what the customer wants, what the company delivers, and what should be billed—dramatically reducing manual work and rework by optimizing the end-to-end process.
Can you talk about other straightforward ways that telcos can use AI to improve performance?
AI in customer service is increasingly table stakes now; many telcos use AI-enabled chatbots and agent augmentation. The next frontier is product development, by letting AI listen to sales and service interactions. The technology can help pinpoint the root cause of problems and scan trends for potential new features, such as adding fraud protection for certain retail customers, and inform upstream product development decisions. By feeding that insight into product and design, AI evolves from a pure cost lever to fuel new growth as well. You’re not just solving problems—you’re identifying new opportunities and improving margins at the same time.
How can telcos implement AI in a way that increases their odds of success?
Focus on five critical elements. First, set up a cross-functional team that is led by business but includes participants from IT, data, and digital. Provide the team with authority over the entire end-to-end process and make it accountable for end-to-end results.
Second, use a single KPI, value realized, to gauge the team’s performance and incentivize them. Ensure that P&L owners will commit resources if the team hits its operational targets.
Third, sequence initiatives for success over different time periods: short-term efforts to achieve quick wins, medium-term to build AI capabilities, and long-term to really rethink the way the company does business.
Fourth, set up a rigorous transformation office that holds all parties to account and provides a single source of truth.
Fifth, build a culture of failing fast. Not everything will be successful, but companies will unlock the creative potential of AI if they’re willing to embrace the technology, identify issues early, and build experience along the way.
