One of the most important questions in technology strategy has always been deceptively simple: who owns the customer? In practice, this is not about legal ownership. It is about who is trusted to help the customer make a decision, who controls discovery and who can influence what happens next.
Search engines, app stores, marketplaces and social platforms built extraordinary businesses by occupying those influence points. They reduced friction for users and, in doing so, gained considerable influence over the companies seeking to reach them.
Artificial intelligence is now changing the character of that interface. Rather than searching, comparing, clicking and transacting across different services, customers may increasingly ask an intelligent agent to do the work. This is an attractive proposition for consumers — but an unsettling one for the companies whose brands, websites and apps the agent may never need to display.
From destination to delegation
The internet economy was built around destinations: a retailer’s website, a bank’s app, a streaming platform, a telco’s customer portal. Distribution strategies followed a familiar logic: attract attention, convert a visit, retain the customer and build direct relationships.
An agent-based economy works differently. A user might instruct an assistant to find the best mobile plan, manage a streaming bundle, book a holiday or resolve a service problem. The assistant can compare options and potentially act on behalf of the user. The destination becomes less important than the quality of the underlying service and the terms on which an agent can access it.
This does not mean every website or app disappears. People will still want entertainment, reassurance and human connection. But where the task is transactional, the interface could move one step further away from the provider.
Where does the value migrate?
- The trusted agent. The system that understands the customer’s preferences and has permission to act may own the most valuable point of influence.
- The underlying service. The best product, price, reliability or content still matters, but it may increasingly need to be intelligible and accessible to machines.
- The data and permissions layer. Identity, authorisation, payments, security and interoperability become strategic infrastructure.
- The scarce asset. Exclusive content, differentiated experiences, physical networks and trusted brands can retain bargaining power even as interfaces change.
The central strategic risk is disintermediation. A business can continue to fulfil transactions and generate revenue, yet gradually lose visibility into why customers choose it and the power to influence that choice.
Why media and telecommunications should pay attention
For media companies, the implications reach beyond search referrals. If discovery and recommendations shift to AI systems, publishers and platforms need to revisit the relationship between content value, distribution and attribution. The commercial question becomes: what is the economic return from having your content used, summarised or recommended by an agent?
For telecommunications operators, the issue is similarly broad. If an external assistant can compare and switch plans or manage a family’s connectivity, traditional digital channels may lose some of their relevance. On the other hand, operators could use trusted access, customer context and service execution to make themselves indispensable parts of the new experience.
A practical response
I would not begin with a grand ‘AI channel strategy’. I would begin by mapping the moments in which consumers currently make consequential decisions, asking which of those moments could be delegated, and identifying the capabilities needed to win when they are.
Some organisations will choose to build their own assistants. Others will partner with the major AI interfaces. Many will need both. But the objective should not be to defend an app for its own sake. It should be to preserve the economics and trust that make the customer relationship valuable.
In a world of intelligent agents, access to a customer is not the same as influence over the customer. The companies that understand that distinction early will have more options later.
Further reading: Deloitte Global TMT Predictions 2026 explores AI in search, orchestration and the evolution of software interfaces.
Leave a Reply