What is Agent Experience?
Back in the ’90s, Don Norman coined User Experience (UX). We take it for granted today, but the idea that a product isn’t just a checklist of features but that users actually have an experience in the interaction with that product that can differentiate it from other products was novel back then.
Then, in the 2010s, Jeremiah Lee talked about Developer Experience (DX), how developers interact with your platform: your APIs, documentation, CLI, and SDKs.
Now, we’re seeing the rise of a new persona that uses our products, the autonomous agent. Agents aren’t programmed to follow fixed steps; they figure things out dynamically. That means we need to design for them, too.
Agent Experience (AX) is about creating products that agents can navigate, integrate with, and orchestrate effectively.
Open vs. Closed Agent Ecosystems
Some companies are taking a closed approach, building their own proprietary agents and ecosystems. Others are building for openness and interoperability and creating great agent experiences so external agents can integrate easily.
For example, Salesforce is going all-in on its “Agentforce” platform, while HubSpot is creating open agent marketplaces and MCP servers that make it easy for agents to access data and workflows.
The same contrast exists between Asana and Linear. Asana builds its own embedded agents, while Linear focuses on openness, letting agents like ChatGPT or Cursor integrate seamlessly. It’s a philosophical divide between owning the agent layer and optimizing for agent experience.
How is AI Changing Web Development?
The barrier to entry will drop dramatically. It’s already easy to spin up a new site with just a prompt. But at the same time, new layers of complexity will appear: Is your app agent-ready? Does it have an MCP or API endpoint? How does it integrate into AI indexing?
In the short term, AI will make it much easier to build the kinds of web projects we already know. But if you want to stand out, you’ll need to push the boundaries of creativity and complexity again. It’ll be simpler to make ordinary things but harder to make extraordinary ones.
Everyone is a Developer
When we started Netlify, the idea behind Jamstack was to decouple the frontend experience layer of the web from the backend infrastructure and to build real tooling around that. We believed that frontend developers would become mainstream web developers, and that absolutely happened.
Today, there are around 17 million JavaScript developers building with modern frameworks on platforms like ours. Now, AI is driving the next foundational shift. There’s a whole new cohort of “vibe coders” who aren’t traditional developers but are still shipping products. If we can build safe, end-to-end tooling for them, this might just become what we call web development in the future.
I think people still underestimate what it means that LLMs can generate code, and how fast they’re improving. It’s like when writing went from being a skill of the elite to something everyone does daily.
Centuries ago, being a scribe was a profession in itself. Then the printing press democratized writing, and suddenly everyone needed to write.
Now the same thing is happening with code. For a long time, it was binary: you either knew how to code or you didn’t. That’s changing. Soon, every profession will require some ability to work with code, just like everyone today needs to write emails or documents.
In the next 10 years, ideally, everyone would be a web developer, at least to a degree. Not everyone will be a professional software engineer, but it should feel frictionless to go from idea to creation without needing permission or gatekeepers.
The Decade of Agents
We’re at the start of the decade of agents. In 2026, we’ll see real-world impact get to the point where they start going deeper into how we work on them. I think the initial hype of vibe coding will be replaced with the massive actual impact of people starting to realize that they can build software for all kinds of purposes now.
But we'll also realize how much work from a platform level we need to put in place for people to be able to use agents in professional contexts where they're trying to work with existing legacy code bases and data sources. And I'm pretty excited for how that will change people's relationship to not just being consumers of software but creators of software.
Takeaways
- AI is fundamentally changing the coding landscape.
- The line between developers and non-developers is blurring.
- Agent experience (AX) is a new discipline in development.
- Automation is crucial for managing increased code production.
- The future of development will be more fluid and less binary.
- Simplicity in frameworks leads to better developer experiences.
- Open systems may provide more flexibility than closed systems.
- AI is a tool that enhances human creativity.
- The web remains the fastest platform for distribution.
- The decade of agents will redefine how we build software.
Chapters
00:00 Introduction to Developer Experience and AI
03:00 The Evolution of Coding and Development
06:07 The Impact of AI on Software Development
09:59 Understanding Agent Experience (AX)
13:40 The Future of Human and Agent Collaboration
17:50 Navigating Open vs. Closed Systems in Development
21:50 Simplicity in Development Frameworks
25:31 The Role of Agents in Web Development
29:34 Predictions for the Future of Development