Most conversations about AI's impact on business are still framed around productivity: creating faster, automating repetitive work, reducing costs, and helping smaller teams accomplish more.
Those are real benefits, but they describe only the first layer. The more consequential question is economic: what happens to market power when the cost of intelligence, search, prediction, and decision-making begins to fall at the same time?
That is where the impact of AI becomes much more interesting. AI is not simply improving how companies market, sell or operate. It is beginning to change how markets themselves function, particularly who controls information, demand and the decisions that connect the two.
From Search to Delegation
The internet dramatically reduced the cost of finding information. Search engines helped us discover products, marketplaces helped us compare them, and social platforms helped brands compete for attention.
AI agents could take the next step by reducing the cost of making the decision itself.
Instead of researching ten hotels, comparing reviews and checking prices, a customer could increasingly delegate that process to an AI agent. NBER research describes AI agents as systems capable of searching, negotiating, and transacting on behalf of users, potentially reducing search and transaction costs. That creates a new intermediary between supply and demand.
And whenever an intermediary becomes responsible for navigating a market, it can acquire significant influence over where demand goes.
The question therefore changes from: How do we get discovered?
To: How do we get selected?
The Next Gatekeeper Could Be Algorithmic
For decades, companies have competed for shelf space, search rankings, social feeds and marketplace visibility. AI could create a more consequential form of gatekeeping because it may increasingly determine which options enter the customer's consideration set in the first place.
Research published in the Oxford Review of Economic Policy found that recommendation systems can reduce search costs while also increasing market concentration by creating "superstar" products. Its modelling also found that recommendation systems could support significantly higher prices under certain conditions.
There is an important lesson here. We tend to assume that more information automatically creates more competition. It does not necessarily. Better information can strengthen winners when it is filtered through a powerful intermediary.
This could make algorithmic consideration one of the most valuable forms of market access.
Data Could Become Organizational Memory
The next source of competitive advantage may not be AI itself. It may be what a company has accumulated for AI to learn from.
Customer searches, transactions, product usage, complaints, returns and service interactions can become inputs into a continuous learning system. The value is not simply that these signals can be analysed faster. It is that they can improve decisions across the organisation.
That creates a feedback loop:
-
Better customer intelligence improves decisions.
-
Better decisions improve customer experience.
-
Better experiences generate more useful data.
-
More useful data improves the next decision.
The organisation effectively develops institutional memory.
The OECD has identified data feedback loops, economies of scope, and access to critical AI inputs as important competitive considerations in developing AI markets.
This is why the AI business impact should not be measured only through productivity. A company that learns faster than its competitors can create an advantage that is significantly harder to copy than an automation workflow.
The AI Economy Has a Concentration Paradox
Here is the contradiction I find most interesting.
AI can make sophisticated capabilities cheaper and more accessible while simultaneously increasing the importance of highly concentrated resources such as compute, foundation models, specialised talent and infrastructure.
Research in Economic Policy identifies significant economies of scale and scope in foundation-model development, while the OECD has highlighted both the opportunities AI creates for smaller firms and the structural advantages available to companies controlling critical inputs.
So the AI economy may produce two seemingly opposite outcomes at once: Capability becomes more distributed. Infrastructure becomes more concentrated. That distinction will matter enormously because access to technology and control over technology are not the same thing.
What Happens to Brands?
This is where marketing comes back into the picture.
If AI increasingly helps customers evaluate alternatives, brands will need to operate in two environments simultaneously. They must remain meaningful to people, but also provide the signals that allow intelligent systems to understand their relevance, credibility, and value.
That makes several traditionally separate brand assets more strategic:
-
Reputation and trusted third-party signals
-
High-quality product information
-
Customer experience data
-
Distinctive positioning
-
Structured and accessible knowledge
-
Proprietary customer intelligence
Brand could therefore evolve from primarily being an attention asset into becoming a decision asset. A strong brand has always reduced uncertainty. In an AI-mediated market, that function could become even more valuable.
The Question CEOs Should Be Asking
The important question is no longer simply, "How is AI used in business?" That will soon become as ordinary as asking how a company uses cloud technology. The strategic questions are more fundamental:
-
Where is AI changing the economics of our industry?
-
Who is gaining influence over customer decisions?
-
Which of our advantages are becoming easier to replicate?
-
Which assets become more valuable because AI can amplify them?
And perhaps most importantly: Where will the value created by AI actually accumulate?
With customers, through lower search and transaction costs? With platforms, through control of discovery? With brands, through proprietary intelligence? Or with infrastructure providers, through control of the underlying technology?
We do not have definitive answers yet. The competitive effects will vary by industry, business model, and access to critical inputs. But one shift is already worth paying attention to. The internet made information abundant. AI could make decision-making increasingly abundant and increasingly delegated.
When that happens, information may no longer be the scarce resource. The scarce resource may become the ability to influence the decision.
That is why I believe AI will change market power more profoundly than marketing. Businesses thinking seriously about the future of AI in business should look beyond automation and ask a more fundamental question:
As intelligence moves through the market, who gets to influence what gets discovered, considered, trusted, and ultimately chosen?
That is where the next generation of competitive advantage may be built.