Humans Are Fundamentally Reluctant to Choose
In Thinking, Fast and Slow, Nobel Prize-winning psychologist Daniel Kahneman distinguished human cognition into two systems: System 1, which is fast, intuitive, and automatic, and System 2, which is slow, deliberate, and analytical.
Kahneman’s work revealed an uncomfortable truth about human nature: people are fundamentally cognitive misers.
Although the human brain accounts for only about 2 percent of body weight, it consumes roughly 20 percent of the body’s energy. As a result, people naturally avoid activating System 2, which requires conscious effort and significant cognitive resources.
Yet modern life forces us to make countless small decisions every day.
Should I eat at McDonald’s or Burger King?
Which mobile carrier offers slightly lower fees?
Should I fly JAL or ANA, and which hotel combination is optimal for my business trip?
Among Amazon Prime and Netflix, what should I watch tonight?
In The Paradox of Choice, psychologist Barry Schwartz argued that an abundance of choices does not necessarily increase happiness. On the contrary, excessive options often reduce satisfaction, increase regret, and make decision-making more difficult.
Similarly, researchers including Roy Baumeister introduced the concept of decision fatigue, describing how repeated decision-making depletes mental resources.
Modern consumers are exhausted. Surrounded by overwhelming amounts of information, they constantly worry that a better option may exist somewhere else. They spend their days searching, comparing, and second-guessing.
This creates the perfect environment for the explosive adoption of AI agents capable of fully outsourcing decision-making.
Traditional information technology merely amplified choices. It presented options and forced humans to decide. The next generation of AI goes much further: it predicts on behalf of humans, decides on behalf of humans, and acts on behalf of humans.
Beyond Prediction: From Prediction Machines to Decision Machines
In their influential 2018 book Prediction Machines, professors Ajay Agrawal, Joshua Gans, and Avi Goldfarb argued that the essential economic value of AI lies in dramatically reducing the cost of prediction.
According to their framework, decision-making consists of five elements:
- Data
- Prediction
- Judgment
- Action
- Outcome
Data serves as the input.
Prediction estimates missing information about the future.
Judgment evaluates the desirability or utility of different outcomes.
Action implements the chosen decision.
Outcome reflects the resulting consequences.
The authors argued that although AI could dramatically reduce prediction costs, humans would continue to perform judgment because only humans could determine values and priorities.
However, the emergence of AI agents is already rewriting this assumption.
AI is evolving from a prediction machine into a system capable of simulating human preferences, making judgments, and autonomously taking action.
Traditional AI (2018)
Data → AI Prediction → Human Judgment → Human Action
Humans remained at the center of decision-making.
AI Agents (2026)
Data → AI Prediction + AI-Simulated Judgment + Autonomous Action
AI increasingly completes the entire decision process.
Why AI Can Now Perform Human Judgment
The major difference between 2018 and today lies in the advancement of large language models and personalization technologies.
Modern AI systems can increasingly approximate an individual’s utility function—their preferences, priorities, trade-offs, and risk tolerance.
Your personal AI agent may eventually understand:
- Your purchasing history.
- Calendar and travel behavior.
- Social media activities.
- Health and sleep data from wearable devices.
- Financial habits.
- Lifestyle patterns accumulated over many years.
As a result, AI may understand preferences that individuals themselves cannot fully articulate.
When AI becomes superior at both prediction and preference modeling, humans have little incentive to activate System 2 and personally intervene in routine decisions.
The human role gradually shifts from making thousands of micro-decisions toward governing the overall behavior and objectives of one’s AI agents.
Why Consumers Will Voluntarily Delegate Decisions to AI
Consumers do not outsource decisions merely because AI is intelligent.
They outsource because doing so increases overall utility.
Even if AI occasionally chooses an option slightly inferior to what a highly engaged consumer might have selected personally, two benefits frequently outweigh this gap:
- Lower cost.
- Near-zero cognitive effort.
Spending two hours searching for the perfect option often generates less total utility than instantly receiving a sufficiently good solution with no effort.
In economic terms, reduced search costs and reduced cognitive costs increase net utility.
People will therefore willingly delegate decisions to AI not because they become incapable of choosing, but because choosing becomes economically irrational.
The Decline of Advertising
This transformation fundamentally changes competition.
Historically, companies fought for human attention.
They invested in advertising, packaging, branding, and emotional storytelling.
But what happens when AI makes the purchasing decision?
AI does not respond to television commercials.
AI is largely indifferent to attractive packaging.
Instead, AI evaluates:
- Price
- Performance
- Reliability
- Historical outcomes
- Reputation data
- Quality consistency
Competition shifts from impressions to verifiable facts.
This does not mean brands disappear.
Brands evolve.
Traditional brands generated emotional affinity.
Future brands increasingly function as trust infrastructure.
When products offer similar prices and performance, AI will favor companies with consistent quality, fewer failures, and stronger historical reliability.
Brand equity becomes accumulated trust data.
Whose Interests Does AI Represent?
A critical question emerges.
Who does the AI work for?
If AI systems are financed primarily through advertising, they may optimize for advertisers.
If users directly pay for AI services, agents are more likely to optimize for consumer interests.
The answer will shape future market structures.
Multiple models will likely coexist for some time, and the battle over AI alignment may become one of the defining economic issues of the coming decade.
Avoiding Commoditization in the API Economy
Even highly sophisticated APIs ultimately risk price competition.
To preserve pricing power, companies must develop forms of value that cannot be replicated.
As more economic activities become digital, the value of non-replicable assets rises.
Drawing on the theory of dynamic capabilities developed by economist David Teece, companies must transform their asset portfolios from easily imitated brand images toward difficult-to-copy physical, intellectual, and experiential assets.
In hospitality, this may mean:
- Unique locations with irreplaceable views.
- Authentic cultural experiences.
- Exceptional craftsmanship.
In manufacturing, it may involve:
- Proprietary materials.
- Patent-protected technologies.
- Exclusive production capabilities.
- Highly specialized equipment.
If no alternative exists, AI agents will select the supplier regardless of price.
The Polarization of the API Economy
The emerging economy increasingly divides into two layers.
Commodity Layer
Companies offering standardized products with no unique advantage face relentless price competition. AI systems continuously compress margins toward zero.
Unique Asset Layer
Companies possessing irreplaceable assets, technologies, experiences, or capabilities retain pricing power and strategic control.
The API economy therefore eliminates the middle ground.
Markets polarize between extreme efficiency and extreme uniqueness.
Implications for CEOs
What should business leaders do?
The answer is clear.
Build unique assets and optimize not only for human perception but also for AI evaluation.
Three requirements become essential:
- Express product value in AI-readable data.
- Ensure that data is trustworthy and verifiable.
- Make information accessible through APIs and external systems.
Only then can a company become part of AI-driven decision-making.
Of course, the transition will not occur overnight.
For many years, markets will remain hybrid environments in which humans and AI coexist.
Companies must therefore compete on two fronts simultaneously:
- Experiences and emotions for humans.
- Data and performance for AI.
Managing both dimensions will be extraordinarily difficult.
Yet only those organizations that successfully navigate this transition will remain among the chosen companies of the next era.
Consumers are gradually ceasing to be the primary decision-makers.
Decision authority is quietly but steadily shifting toward AI agents.
The defining question for every company becomes:
How do we become the company that AI chooses?
The answer to that question may determine competitive advantage in the age of AI.
To be continued.
Platform Strategy® is a registered trademark of NetStrategy Inc.
