What if the future of design isn’t about interfaces, screens, or even journeys — but about intents?
Not tasks.
Not steps.
Not features.
Just: What does the user want to happen?
Lately I’ve been noticing a pattern across AI, product design, crypto, and even my own daily life. Everything seems to be collapsing into a single primitive: intent. Declare the outcome, and let the system handle the rest.
The Mental Models Behind This
I’m drawing on several frameworks to explore this:
- First Principles Thinking – Stripping interfaces down to their fundamental purpose: revealing user intent
- Jobs-to-be-Done (Clayton Christensen) – Users don’t want products, they want progress. Intent is the purest expression of desired progress.
- Systems Thinking – Understanding how complex intents decompose into executable actions
- Disruptive Innovation (Clayton Christensen) – Intent-based systems start simple but will move upmarket to displace complex interfaces entirely
The insight comes from combining these: in technological transitions, the winning approach is often the one that strips away the most unnecessary complexity.
From “How” to “What”
For decades, technology has forced us to think in procedures:
- Search for flights, compare dates, choose seats → buy a ticket
- Scroll, filter, check stock → find a gift
- Click through menus → relax
But something big is shifting. Tools like ChatGPT, Gemini, and Midjourney have flipped the model: you tell them what you want, not how to get there.
Even in crypto, Uniswap’s ERC-7683 intent standard for cross-chain trading lets you specify the end state — “swap X for Y at the best price” — and a network of solvers competes to fulfill it. Commerce is moving the same way: “Tell me when size 10s drop under €100,” or “Find me a present for a nephew who loves dinosaurs.”
The user provides intent. The system figures out the rest.
This shift sounds simple, but it changes everything.
Agents: The Accelerant
Here’s the part that makes this moment feel urgent:
It won’t just be humans expressing intents — it’ll be AI agents acting for them.
- “My user wants office supplies by Monday — optimize for sustainability”
- “My user wants the cheapest contract-free phone plan — negotiate”
- “My user wants a holiday in June — book something they’ll love”
When millions of agents start firing intents at the world simultaneously, we might see market dynamics humans have never experienced: machine-speed commerce, negotiation, planning, logistics, content flows.
Platforms that aren’t designed to receive, interpret, and fulfill intents could feel suddenly… dated.
Intents as the Design Primitive
Once you see it, you can’t unsee it:
Every interface is just a clumsy way of getting you to reveal your intent.
- A search bar is a primitive intent slot
- Filters are intent crutches
- A checkout flow is intent translation
- A prompt is explicit intent engineering
Prompt engineering is intent engineering. It’s learning to express a future state (“Generate a logo…”, “Rewrite this as…”, “Find flights…”) so the system can navigate there.
This mirrors Jeff Hawkins‘ Thousand Brains theory: intelligence emerges from maps, destinations, and recursive sub-goals. Intent is the connecting tissue — the relationship between where I am and where I want to be.
Our technology is starting to look the same: nested intents decomposed until they hit executable actions.
It really is intents all the way down.
Personal, Not Just Digital
This isn’t only about tech. Think of your own life:
- “My intent is to relax”
- “My intent is to be a better dad”
- “My intent is to stay focused today”
These are real intents — vague, high-level, meaningful. If technology understood them, it could support us in ways that feel almost human:
- Relaxing → adjust lighting, quiet notifications, suggest a film
- Being a better dad → surface weekend ideas, manage logistics, remember events
- Staying focused → shape your environment, triage incoming requests
The line between “personal growth” and “product design” gets thinner when both are fundamentally about understanding intent.
Content for Fun vs. Utility
This connects to another pattern I’ve been exploring: the intent of consumption matters as much as the content itself.
When you’re browsing for fun, you want serendipity, discovery, rabbit holes. When you’re trying to accomplish something, you want precision, speed, and signal.
The same piece of content serves different intents depending on context. A cooking video might be entertainment while you eat lunch, or critical reference while you’re mid-recipe.
Understanding user intent transforms how we design, organize, and surface content.
The Architecture Beneath
If intent becomes the new interface, the architecture underneath must change:
1. Intent Recognition
Systems need to understand what the user means, not just what they say.
This isn’t just NLP—it’s inferring context, learning patterns, understanding the difference between “I want a gift” (exploration) and “I need a gift by Friday” (urgency).
2. Capability Maps
Merchants, apps, APIs, and agents must advertise what intents they can fulfill.
Imagine a world where a store’s competitive advantage is:
“We fulfill urgent intents better than anyone else.”
3. Fulfillment Orchestration
Intents become routable objects — split, delegated, negotiated, or bid on.
Your “I need dinner” intent might decompose into:
- Find nearby restaurants with availability
- Filter by dietary preferences
- Optimize for time or cost
- Reserve table
- Navigate there
All handled, not by you clicking through six apps, but by the system understanding and executing your intent.
4. Trust and Constraints
This is where it gets interesting:
- “My intent is private”
- “Don’t optimize for cost at the expense of ethics”
- “Never auto-commit purchases over €200”
We’ll need rule systems that respect meta-intents: the intents governing our other intents.
AI as Question Machine, Not Answer Machine
Here’s a related idea: what if AI’s real value isn’t giving us answers, but helping us understand our own intents?
Like the Hitchhiker’s Guide: the answer to life, the universe, and everything is “42” — but the real insight is that we never properly understood the question.
AI that coaches us toward clarity (“What are you actually trying to achieve here?”) might be more valuable than AI that just executes blindly.
There’s a balance:
- Eliminate bad friction (tedious steps, unnecessary complexity)
- Create good friction (pause points, reflection, understanding)
The best intent-driven systems will know when to get out of the way, and when to slow you down to make sure you actually want what you’re asking for.
Bret Victor and the Principle of Intent
This reminds me of Bret Victor’s “Inventing on Principle” talk, where he demonstrates systems that respond immediately to your actions, letting you see consequences in real-time.
Maybe there’s a new principle here:
“Help users understand and achieve their intent.”
Not:
- ❌ “Make it easy to use our features”
- ❌ “Guide users through our flow”
- ❌ “Optimize conversion”
But:
- ✅ “Understand what the user wants to happen, and make it happen”
This shifts the entire design question from “How do we expose our capabilities?” to “How do we understand their intentions?”
Why This Matters Now
This feels like more than just a UX trend—it might be a paradigm shift:
- AI is pushing us toward outcome-driven interactions
- Crypto is reorganizing systems around intent-based settlement
- Commerce is moving from catalogue-push to intent-pull
- Agents will soon operate at a scale where procedural interfaces collapse
What we design next needs to assume:
The user will express intent directly. The system must do the rest.
Where I’m Going with This
This piece is the start of an exploration. Over the next few articles I want to dig into:
- How agentic commerce flips the whole e-commerce model
- How intent-based systems change incentives and business models
- How privacy, identity, and consent work in an intent-driven world
- How to design UIs for people, and APIs for agents
- Why “meta-intents” might matter more than intents themselves
If the last era of the web was about content, and the current one is about context, the next one appears to be about intent — the cleanest expression of what a human wants.
The Thousand Brains Connection
Jeff Hawkins’ Thousand Brains theory, developed at Numenta and detailed in his book “A Thousand Brains: A New Theory of Intelligence,” proposes that intelligence emerges from thousands of cortical columns, each building models of the world through reference frames—essentially maps with locations and movements.
The key insight: Intelligence is about knowing where you are and where you want to be, then navigating between them.
Sound familiar?
That’s intent.
Our brains work by:
- Building models (maps of possibility)
- Identifying current state (where am I?)
- Identifying desired state (where do I want to be?)
- Computing paths (how do I get there?)
This recursive process happens at every level—from moving your hand to planning your career. Each level deals in intents that decompose into sub-intents until you hit executable actions (motor commands, API calls, purchases).
Our technology is evolving to mirror this architecture.
Where I Might Be Wrong
This analysis assumes a few things that could turn out to be incorrect:
I’m assuming that intent recognition will become reliable enough to trust. If the technology for understanding what users actually want remains clumsy or error-prone, we might be stuck with explicit interfaces longer than I’m projecting. Privacy concerns could also limit how much context systems can use to understand intent.
I’m assuming users will want to cede control to intent-driven systems. But there’s real value in procedural interfaces—they create understanding of how systems work. If intent-driven design reduces user agency or creates new forms of manipulation, adoption could be much slower than I’m modeling.
I’m assuming that the displacement follows Christensen’s disruption playbook—starting simple and moving upmarket. But procedural interfaces might persist in domains where explainability matters more than convenience (healthcare, finance, legal).
An alternative view would emphasize the trade-offs: intent-driven systems might reduce serendipity, create filter bubbles, or eliminate the productive struggle of figuring things out yourself. The best outcome might not be pure intent-driven design, but a hybrid that knows when to get out of the way and when to engage users in the process.
I could be wrong about the timeline, the adoption curve, or whether this is truly a paradigm shift versus just another interface pattern. But if the pattern is real, the implications for product design are significant.
Conclusion: All the Way Down
When you zoom out far enough, it becomes obvious:
It’s intents all the way down.
From the moment you wake up (“My intent is to feel rested”) to the apps you open (“My intent is to connect”) to the purchases you make (“My intent is to solve this problem”) — everything is intent.
The interfaces we’ve built—buttons, forms, menus, flows—are just scaffolding around this fundamental truth. As AI gets better at understanding what we want, that scaffolding becomes unnecessary.
The future of technology isn’t about better buttons, it’s about better understanding.
And when the system truly understands your intent—when it can decompose it, route it, fulfill it, and adapt based on feedback—the interface disappears entirely.
What remains is pure expression of will. The work? Understanding what your intents are.
Intent → Outcome.
Nothing in between.
Published: December 9, 2024
Related Content:
- Intent-Driven Commerce: What E-commerce Can Learn from AI Agents and DeFi – Technical implementation of intent-driven systems
- Web Summit 2025: Orchestration, Flow, and Dynamic Resource Allocation – Intent not instructions theme
- There Is No Host: Freedom Stacks – Operationalizing freedom through technology
Referenced Concepts:
- A Thousand Brains: A New Theory of Intelligence by Jeff Hawkins – Numenta research
- Bret Victor’s “Inventing on Principle” – Classic talk on immediate feedback
- Uniswap Labs and Across Propose ERC-7683 Standard – Cross-chain intent standard
- Clayton Christensen’s Jobs-to-be-Done – Harvard Business Review
- The Innovator’s Dilemma by Clayton Christensen
- First Principles Thinking – Farnam Street
- The Intention Economy – Doc Searles