Satya Nadella predicts AI agents will disrupt web advertising models
Microsoft's CEO argues that the shift toward AI agents could force a transition from ad-based traffic to transactional commerce.
The short version
- AI agents may shift internet economics from ad-based 'eyeball' models to transactional wins.
- Nadella compares the current shift to the historical move from web portals like Yahoo to search engines like Google.
- Amazon is currently blocking AI agents from browsing its catalog or making purchases for users.
- Platforms like Shopify could gain a competitive edge by allowing agents to interact with their services.
Why it matters
This shift could fundamentally change how companies make money online, moving away from showing ads to people toward facilitating automated transactions. For consumers, it means a future where AI handles the logistics of shopping and browsing on our behalf.
The end of the human web
Microsoft CEO Satya Nadella recently shared his perspective on how artificial intelligence will fundamentally alter the internet's economic structure. He argues that we are moving toward an 'agentic web' where AI agents perform tasks instead of humans browsing pages. This transition poses a significant challenge to the current business models of many major tech companies.
Currently, much of the internet relies on what Nadella calls the 'human web.' This is a system designed for people to click on links, view content, and interact with advertisements. In this model, companies make money by capturing and holding the attention of human users for as long as possible.
Nadella suggests that this era is beginning to fade as AI agents become more capable. If an AI is doing the browsing and the buying, the traditional way of monetizing that activity must change. He believes the current aggregators who built their wealth on human traffic will see their positions reshuffled.
From ads to transactions
A major part of this economic shift involves how companies will be paid for traffic. Nadella notes that if a business currently relies on advertising because of human 'eyeballs,' it may need to pivot toward a more transactional approach. This means the focus will shift from how many people see a product to how many successful actions the AI completes.
He specifically mentioned the possibility of moving toward cost-per-click (CPC) wins. In a CPC model, advertisers pay for every click an agent makes on a specific link. This is a more direct way of measuring success in a world where the user is an automated tool rather than a person looking at a screen.
This change reflects a move toward a more functional internet. Instead of a company trying to keep you on a page to show you more ads, the goal becomes completing a specific task as efficiently as possible. The economic value moves from the attention of the user to the completion of the transaction by the agent.
A historical parallel
To help explain this transition, Nadella compared the rise of AI agents to the historical shift from web portals to search engines. Years ago, companies like Yahoo served as the primary gateways to the internet, hosting various services and links. However, Google eventually disrupted that model by focusing on a more efficient way to find information.
The move from Yahoo to Google was a shift in how users accessed the web. Nadella sees the move to an agentic web as a similar evolution. Just as search engines replaced portals as the primary way to navigate the internet, AI agents may replace manual browsing as the primary way to interact with online services.
This comparison highlights that the current internet infrastructure is not permanent. Technologies that dominate the current era can be replaced by more efficient tools. Companies that fail to recognize this shift risk becoming the next generation of 'portals' that are eventually bypassed by newer technologies.
The battle over access
The practical implications of this shift are already appearing in how major retailers handle AI. As reported by TechCrunch, Amazon has made the decision to block AI agents from accessing its retail website. This restriction prevents agents from browsing the company's catalog or making purchases on behalf of their customers.
Nadella suggests that this defensive move might backfire in the long run. By blocking agents, Amazon is trying to protect its existing business model which relies on human interaction. However, he argues that if a company is too in love with a specific model, they may find themselves struggling to adapt as the market moves forward.
This creates a potential opening for competitors who are more willing to embrace the technology. For example, Nadella suggested that a platform like Shopify could gain an advantage by allowing AI agents to interact with its services. If Shopify makes it easy for agents to shop, it could capture the business that Amazon is currently trying to gatekeep.
Scrambling to adapt
The overarching theme of Nadella's prediction is the necessity of adaptation. He believes that the companies that survive will be those that accept the presence of AI agents in the ecosystem. If a business refuses to allow agents to interact with its products, it risks being left behind by more flexible competitors.
The tech executive noted that this is not a unique problem for the internet. He observed that similar shifts are happening across many different categories of business. When a fundamental technology changes how a service is delivered, the old way of doing things often becomes obsolete very quickly.
For the reader, this means the way we shop and find information will likely become more automated. Companies will have to decide whether to build walls against AI or build bridges for them. The winners of this transition will likely be the ones who prioritize the agent's ability to complete a task over the human's ability to view an ad.
What to watch
- Whether other major retailers follow Amazon's lead in blocking AI agents.
- How Shopify and other platforms adjust their APIs to accommodate agent-based shopping.
- The development of new 'transactional' ad models that focus on agent actions rather than human views.
Try it
Look at your favorite shopping site and see if you can find any 'buy on behalf of' features or if they have clear rules against using automated tools.
Sources and links
- ReportingTechCrunch: the article we sourced
First reported by TechCrunch. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.