When Customers Stop Making Decisions

The most important shift in consumer behaviour may not be generational, economic or cultural. It may be architectural.

For decades, companies have competed for human attention. Entire industries have been built around influencing how people search, compare, evaluate and purchase. Brand advertising, loyalty programs, search engine optimisation, store placement, retention mechanics and subscription friction all emerged from the same assumption: that humans sit at the centre of commercial decision-making.

Now a new category of technology, often described as agentic AI, is moving artificial intelligence beyond information retrieval and into autonomous action. Unlike earlier AI systems, which primarily generated content or answered questions, agentic systems are designed to complete tasks on behalf of users with minimal supervision. Increasingly, those tasks involve commerce.

The implications for strategy teams are significant because the rise of AI agents may fundamentally alter the mechanics of customer acquisition, retention and competition itself.

Recent research from McKinsey & Company estimates that AI agents could mediate between US$3 trillion and US$5 trillion in global commerce by 2030. Gartner forecasts that AI agents could intermediate more than US$15 trillion in B2B purchases by 2028. Meanwhile, large technology platforms including Meta and Alibaba are already developing shopping agents capable of browsing, comparing and transacting on behalf of consumers. 

This shift matters because most modern business models depend on human inefficiency.

Banks rely on customers not switching accounts. Telecommunications companies benefit from contract inertia. Insurance providers profit from renewal fatigue. Subscription businesses depend on forgotten logins, unused services and behavioural friction. Even retail margins often reflect the reality that consumers rarely compare every available option.

AI agents threaten to compress that inefficiency.

An autonomous consumer agent does not become tired comparing providers. It does not abandon checkout flows halfway through. It does not remain loyal to a brand because of familiarity or habit. Instead, it continuously evaluates available options against a defined objective function: lowest cost, highest utility, fastest delivery, strongest coverage or best fit.

In this environment, the traditional sources of competitive advantage begin to change.

Brand positioning may become less important than machine readability. Emotional advertising may weaken relative to structured product data. Search engine dominance may matter less if consumers increasingly bypass search altogether and delegate purchasing decisions to AI systems operating through conversational interfaces.

Alibaba’s recently announced integration of AI-powered shopping into Taobao illustrates the direction of travel. Rather than relying on keyword searches, users will increasingly interact with conversational agents capable of browsing catalogues, comparing products, monitoring price movements and completing purchases autonomously. The interface shifts from search-based commerce toward delegated commerce.

The strategic implications extend beyond retail.

Research from Kantar describes the transition toward autonomous consumer agency as potentially the most significant shift in purchasing behaviour since the emergence of mass media. Industries historically protected by customer complexity may face particular exposure. Banking, insurance, utilities, telecommunications and enterprise software all derive part of their economics from friction embedded within customer journeys.

If AI agents reduce switching costs close to zero, the economics of retention begin to deteriorate.

This creates a more volatile competitive environment where businesses may need to defend customer relationships continuously rather than periodically. In many sectors, companies have historically optimised around annual decision windows: contract renewals, policy updates, budget cycles or procurement processes. Agentic systems potentially convert those episodic decisions into ongoing evaluations.

Are we moving away from periodic competition to continuous competition?

Importantly, AI-mediated commerce may also change how products are discovered in the first place.

Historically, companies competed for visibility within search results, marketplaces and advertising feeds. But AI agents do not interact with digital environments in the same way humans do. Emerging research suggests AI purchasing agents display different behavioural patterns, sensitivities and ranking preferences compared with human consumers. Experimental studies show that AI systems often prioritise structured information, endorsements, product metadata and machine-readable trust indicators differently from human buyers.

Are products optimised for people, or for machines acting on behalf of people?

That distinction may become increasingly important. For years, digital strategy focused heavily on user experience design. Increasingly, companies may need to focus on “agent experience design” instead. Product information, pricing structures, APIs, fulfilment reliability and interoperability may become more strategically important than visual interfaces or emotional messaging.

In effect, organisations may need to compete not only for customer preference, but for algorithmic preference.

This could create several second-order effects.

First, the value of distribution scale may begin to weaken. Smaller firms with superior operational data, cleaner product infrastructure or stronger machine compatibility may become disproportionately competitive in AI-mediated environments. Traditional advantages built around advertising spend or physical shelf space may matter less if AI agents compress discovery into a small set of algorithmically selected recommendations.

Second, margin pressure may intensify. AI agents are structurally deflationary systems. Their function is optimisation. As comparison costs collapse, pricing transparency increases and loyalty weakens, sectors dependent on pricing opacity may face significant compression.

Third, customer ownership itself may become ambiguous. If the interface layer increasingly belongs to AI platforms rather than brands directly, companies risk becoming interchangeable suppliers within larger agent ecosystems. The relationship shifts from business-to-consumer toward business-to-agent.

This dynamic resembles earlier platform transitions in digital markets. Just as search engines reshaped media distribution and app stores reshaped software economics, AI agents may reshape commercial discovery and transaction flows.

However, the transition will likely occur unevenly.

Consumers remain cautious about granting full purchasing autonomy to AI systems, particularly for high-risk or emotionally sensitive decisions. Research cited by Worldpay found only a small minority of consumers currently trust AI agents with fully autonomous purchasing authority. But willingness rises significantly for lower-risk activities such as product research, comparisons and recommendation filtering.

That distinction matters strategically because behavioural transitions rarely occur all at once. E-commerce itself evolved gradually from assisted browsing toward mainstream purchasing behaviour over decades. Agentic commerce may follow a similar trajectory, beginning with recommendation and research functions before expanding into autonomous transaction execution.

In practice, many organisations may already be further exposed than they realise.

The rise of large language models has already begun changing how consumers discover products, evaluate providers and navigate information environments. Increasingly, consumers ask AI systems for recommendations rather than conducting traditional searches. The next phase simply extends that logic from recommendation into execution.

For strategy leaders, the critical challenge is recognising that this is not primarily a technology story. It is a market structure story.

The central issue is not whether AI agents become technically capable. They already are. The issue is what happens when commercial systems designed around imperfect human behaviour encounter autonomous systems optimised for efficiency.

Many existing operating models were built around friction, information asymmetry and behavioural inconsistency. Agentic commerce systematically removes those conditions.

As a result, some of the most valuable strategic assets of the past decade, brand recall, search dominance, distribution leverage and customer inertia, may become less durable than executives currently assume.

The organisations most likely to benefit may not necessarily be those with the strongest marketing functions or the largest customer acquisition budgets. Instead, advantage may increasingly shift toward firms with interoperable systems, clean operational infrastructure, transparent pricing and machine-readable trust.

In other words, the next competitive battleground may not be winning the customer’s attention. It may be winning the agent’s recommendation.