The Decision Economy
As AI evolves from passive tools to active agents, we are shifting from an Attention Economy to a Decision Economy where algorithms choose for us.
The first two decades of the 2000s were shaped by the attention economy.
Now I propose that agents will shape a new and more powerful economy. What I label the Decision Economy.
The Roots
Humans have historically been designed for communication, safety, and a sense of belonging.
Our togetherness and ability to come together while also being able to unite under ideas is what gave us an evolutionary advantage.
Technology has made this more apparent: communities, network effects through social platforms, the ability to create status, belonging, niche interests, long-distance communication, open/create new social groups outside of limited environment constraints, attract new mates.
Through this, we were able to capture attention. Primarily through focused areas and groups where there's something to concentrate around. Ideas, beliefs, knowledge, curiosities, and more.
It became more obvious in the 2010s and 2020s how powerful algorithms can now direct this attention towards something.
Right now, our attention is primarily controlled through media, which has been one of the most powerful and significant inventions to mankind.
Since the printing press became a thing, whoever controlled what ideas get stuffed into it and who can sponsor media into the press, controlled the attention currency.
And if you're in socialistic, totalitarian, or more aggressive political countries/groups like a dictatorship, you had very little to no control over what gets seen or communicated inside your bubble of influence (who controls your attention) versus what is out in the "real world."
Then came along the internet. Another powerful breakthrough.
And what's more: mobile apps where news and media can be communicated digitally and decentralized from central sources where friends, peers, influencers, thought leaders, media channels, and more can now compete over attention and who gets seen/heard versus who doesn't.
In the beginning, connection and getting users/people to stay and use the platform was the initial primary metric to optimize for.
For instance, Facebook initially wanted a way to grow their platform fast, retain users, and eventually monetize all the attention they held.
They discovered that if you join the platform and add 7 friends within 10 days, you are far more likely to use and stay using Facebook.
This insight, driven by Chamath Palihapitiya's growth team, became their "North Star" metric and helped propel them from 5,000 daily signups to 70,000.
Connection eventually transformed once platforms became public.
From early 2010s feeds which were of your friends, to now powerful and trained algorithms over billions of human data points and other information from its user base. Algorithms designed to present and show you information/media which is most likely to keep you on the platform and monetize.
I constantly find myself nowadays, after Instagram made their algo feed transparent, looking at what the algorithm feed keywords are and removing/adding what I want to see.
Week over week, I notice that my feed changes/updates to what my behavior reflects versus what I realistically and healthily want to see.
Their algorithm knows what will keep me versus what I want, even if it may cost them how much time I stay on their platform.
Which leads us to here.
The Shift
Now the shift is entering towards the Decision Economy.
An economy where what's online and presented to you is no longer "just" controlled by an algorithm, but by an intelligent system which chooses and decides to show content, information, links, references, sources, and more at its own disposal.
The key difference: traditional algorithms optimize for engagement metrics. They learn what keeps you scrolling and serve more of that.
AI agents go further. They don't just optimize. They decide. They select, filter, synthesize, and present information based on goals that may or may not align with yours.
The Decision Economy works by AI agents being able to control what is shown to you.
Right now, we see it working in infant ways, yet still powerful:
- What information AI is trained on
- How an AI is trained
- What an AI can show you versus what it cannot because it is "against a policy," even if it's a "gray" area which can be considered undefined but normal to speak about
- Whether you're a certain age range, gender, or have personalized context/memory to make it feel more personalized to you
We've already seen a preview of this with traditional algorithms. Look at TikTok.
What China sees versus what TikTok US sees is not the same.
Tristan Harris, a former Google employee and social media ethics advocate, described this difference on 60 Minutes: "It's almost like they recognize that technology is influencing kids' development, and they make their domestic version a spinach version of TikTok, while they ship the opium version to the rest of the world."
That's still an algorithm making those choices. Now imagine that same dynamic, but with AI systems that can reason, contextualize, and make real-time decisions about what you see and why.
We are creating tools which can ultimately control what or how we think. Which I suppose is another large thread I can cover but perhaps will save for another post.
How It Works
To understand the Decision Economy, you need to understand how these AI systems actually operate.
Initially, these chat models were just that: a chat model. You ask, it responds.
As the tools became more advanced, larger in parameters/training sets, and more technically capable (which requires more power/compute to run), they require more sophisticated architecture to operate.
They're more than just hooking up a simple generative LLM to a chat interface.
They are now orchestrators with tools. Multimodal systems which can now intelligently choose what model to select (such as Grok 4's Fast vs. Expert modes) or what modality to enhance/enrich the chat response.
Gemini 3 knowing whether or not to use Nano Banana for image generation, or Veo 3.1 for video generation, or simply provide a text response.
At a high level, these systems now have "orchestrators" equivalent to what operas have with a conductor to coordinate musical labor.
The orchestrator controls and thinks ahead what gets shown, why, and makes an execution decision to perform something for some reason because it believes that's the best decision for a response/piece of output.
This is what I mean by Decision Economy.
Intelligent AI which now controls and decides a smart output for you. Whether it's something you genuinely want or not.
The First Battleground
If AI agents are making decisions about what to show you, then the first major battleground is obvious: search.
Right now, AI search is experiencing explosive growth.
While exact figures vary by source, the trend is undeniable: Perplexity AI queries grew over 500% year-over-year, ChatGPT prompt volume jumped nearly 70% in the first half of 2025 alone, and roughly 50% of Google searches now include AI-generated summaries.
This represents a magnitude of impact similar to when the internet first emerged.
The first to the races of who is trying to "hack" these systems (or should I say, find ways to outcompete or rank over others) is through an emerging space called GEO/AIO. Generative Engine Optimization/AI Optimization. The coming usurper to SEO.
Those who correctly rank, label, and structure their data, webpages, stores, or any other thing they want accessed and used as an asset for some certain reason (sale, reference, context injection for an AI) will win the decisions these AI agents make in terms of who gets shown what and why it should show that over someone else.
Great data structure/credibility = visibility.
Poor data structure/credibility = obscurity.
It's still an undefined space as there's a lot of things people are learning and need to know how to effectively rank/work with these tools.
There's no clear way yet to be certain in your ability to optimize since, after all, these are "generating" models and not absolutely predictable models.
Which is exactly the point. In the attention economy, you could game the algorithm. In the Decision Economy, you're trying to influence an agent that reasons.
The Implications
Now with programmable tools, apps, and other products emerging into society, how much power of choice do we really have?
If an AI has so much context (say purchase history, browser behavior, social media data), it can provide me a response so personalized and tailored to me. And what's more, predicted in advance. That I may jump at the opportunity to get X product because it fits my current pain point like a glove.
In this case, was the tool used for good? Selling me an offer to a pain point which is personalized and perfect to my unique case scenario?
Or is it showing me something it wants me to see over other products/offers/information (maybe even knowledge, right?) that could be useful. But instead it chose to sell me something since the central organization which may own it has an incentive to make its shareholders money and profit.
The Decision Economy is ultimately about how AI decides over what decisions to make and why.
As attention is the currency for humans, decisions will be for AI. Now that they can navigate more and more on their own, soon with little oversight by a human as they become more autonomous.
The Point
The point of this post is not to explain how you can monetize this.
It's instead an exercise in thinking in terms of where society is heading, what to be aware of, and why it makes sense to start paying attention to this sooner.
Because at some point, all the content you can be presented or exposed to on the internet, metaverse, or some other medium where AI has your attention can be completely generated and decided for you instantaneously.
It's more important than ever to understand and form your own thoughts, conclusions, ideas, and beliefs over simply outsourcing them to an AI.
While AI is great as a thinking companion and a great resource (and soon to be perhaps even a friend), you must remember that at the end, we also are biological computers.
Just because there is now an artificial computer capable of making decisions and acting on them, it doesn't mean decisions which are made and controlled for some reason by outside agents should be your default go-to.
AI Agents are here and you cannot do much about it.
What's more is to understand the Decision Economy. It's an important concept to grasp.