Delivering large-scale change in VUCA environments

Ask McKinsey or ask BCG, seventy percent of digital transformations fail to deliver what was expected of them. Preparing for this session I read the risks that the world's top twenty operators publish in their annual reports, and one theme runs through all of them: more volatility, more uncertainty, more complexity, more ambiguity. Transforming is hard already, and in a VUCA world the failure rate gets worse. That raises three questions. How do we set strategy for the medium and long term, how do we set up organisation and delivery, and how do we steer investment and technology toward a return that survives.

Volatility shows up in three places. Demand, as consumers trade down from the main brand to the second brand and enterprise spending follows a macroeconomic environment that has not recovered. Cost and supply, through interest rates, financing costs, energy prices and supply chains. And natural disasters and energy supply, with energy itself named as the coming scarcity in an AI-driven world. At Dialog Axiata we lived this. The Easter Sunday bombing with hundreds of casualties, political instability straight after it, then Covid, which destroyed the tourism earnings the country depended on for foreign exchange. What carried us through was an organisation and a modular, API-based architecture that could react in days: connectivity for families under curfew who could not reach a top-up point, flat rates for educational content when school moved home. Volatility is answered with agility. Build to react faster than the competition.

Uncertainty shows up in another three. Regulation, covering consolidation, spectrum auction rules and supply-chain requirements. Rapid technological change, including whether there is time to monetise 5G before the next generation arrives and how generative AI lands in operations. And investor preferences, because a sector that has delivered poor shareholder returns attracts activists, private equity and pressure to simplify. At Reliance Jio we entered a market that already held eleven carriers, with a new technology, a target of a million customer activations a day, and no reference case anywhere for a greenfield launch at that scale. The regulator had not settled how customers could be acquired, a sudden demonetisation put prepaid recharges at risk, and affordable handsets did not exist. We answered it multi-pronged, so nothing had a single point of failure: two acquisition processes built in parallel, with the identity-based one activated the moment regulation was clear, a payments company so people could recharge online, and our own handset company to get a sub-hundred-dollar 4G device into the market. Uncertainty is answered by hedging. Run a funnel of initiatives and think like a venture capitalist.

Complexity and ambiguity map onto the five Ps. Products, where a legacy portfolio carries declining revenue at high margin. Processes, broken ones producing weak sales, high cost and a customer experience that annoys people. Platforms, where legacy coexists with cloud-native workloads and slicing, across fibre, wireless and satellite. People, with the rise of the freelancer and now the rise of the AI agent, and teams that will hold both. Partners, where co-opetition is real and equipment makers begin serving our customers directly. At Telia we found exactly that complexity, and it showed up as revenue decline, high cost, slow time to market and poor experience. We ran a comprehensive programme to simplify, standardise and scale along the five Ps, and three to four years later we are back to growth on top line, bottom line and customer experience, with legacy down more than half in three years. Complexity is answered by thinking big and executing programmatically.

There is no silver bullet. Transformation never stops and has to become part of a company's DNA, and VUCA conditions, while they make the work harder, hand an advantage to whoever manages them better than the rest.

On simplifying operations, start with the processes and design them end to end, then pick the platforms underneath them. We took ServiceNow as the fabric across IT service management and customer service management, live in every market, which let us correlate a customer complaint with a fault in the machine room. Once a process is digitised you can automate it and then put a generative bot on top. One GitHub for the company, then Copilot. One Jira for collaboration. One data lake and an open AI architecture, with generative AI now in production for a customer service knowledge base, a sales response engine and marketing content.

On portfolio, knowing you have forty thousand products is already a good start, because most of us do not know the number and some do not know what counts as a product. Get clarity, define the target on profitability and strategic value, and then run the migration factory, which is the hard part: stop selling the legacy, take the customer through a stage gate to a target product, and accept that the cost only leaves when the last customer is off the last legacy platform. That takes two to three years.

On talent, we undersell ourselves. Research with technology graduates found their picture of telecoms is a hard hat and a ladder, against a hyperscaler with beanbags. We are a software industry with unusual societal responsibility, deep customer intimacy and a playground for analytics, automation and AI, and we run generative AI at scale today.

On growth, be realistic. This is not a hyper-growth industry and it should not pretend to be. Cash flow matters more than revenue growth in a capex-intense business running at thirteen to fifteen percent of sales. Legacy revenue with high margin is declining and will go. Focus on core connectivity beyond voice, messaging and data, monetise network functions through APIs, and take a few bets close to the core: cloud, security, enterprise mobile networks.