Your Platform Might Already Be Incompatible With the Future

At this year’s Infobip Shift conference, Matt Biilmann Christensen, co-founder and CEO of Netlify, spotlighted the next major evolution in software: agentic experience (AX).

He drew a clear line through past transformations — User Experience (UX) reshaped digital products in the 1990s, and Developer Experience (DX) redefined platforms in the 2010s. Matt argues that AX drives the critical design frontier in the era of AI-driven development.

“Good DX doesn’t always mean good AX,” Matt said. Tools that feel intuitive for humans often frustrate AI agents. He urged builders to design platforms that deliver seamless human usability and strong agentic experience, allowing AI systems to interact, adapt, and create value without friction.

Redefining the Developer Population

Matt opened by challenging how we define a “developer.” He reminded the audience that the industry once dismissed frontend developers as “pseudo-developers,” only later to acknowledge them as essential builders of the modern web.

Similarly, he argued that today’s AI agents sit at the edge of legitimacy. With the right platforms and tooling, they could redefine development and bring billions more people into the fold.

Matt envisioned a future not limited to millions of professional software engineers but open to as many as three billion people — anyone with technical acumen and an internet connection — who could step into development through agents.

The opportunity is not just to grow the developer market, but to redefine what development means for the “next 100 million people coming online”.

Access, Context, and Tools: The Three Pillars of AX

Matt framed AX around three core challenges: access, context, and tools.

Access focuses on how easily agents can interact with a product. Traditional authentication models, like OAuth scopes, often block agents that need open-ended flexibility. Matt highlighted “anonymous flows” and sandboxed environments, which let agents experiment before locking in permissions.

Netlify, Neon, and Clerk provided clear examples of this pattern, granting limited access first and formalizing accounts only when necessary. Context ensures agents understand a product’s existence and how to use it effectively.

Matt spotlighted “context engineering,” where structured documentation — Markdown files, MCP servers, or custom .md rules — feeds LLMs the information they need to generate reliable output. He also suggested that search engine optimization could evolve into Generative AI Optimization, helping models surface and apply product knowledge more accurately.

Tools form the final pillar, shaping how agents act on information. Matt stressed that platforms must offer intuitive APIs, predictable interfaces, and robust debugging options so agents can perform reliably at scale.

For Matt, AX is not a feature or protocol but a discipline, much like UX. It requires constant observation, iteration, and empathy.

On the other hand, tools are the interfaces through which agents interact with products. Even well-designed CLIs and APIs can fail when faced with non-human users.

Matt showed how Netlify’s CLI, long celebrated for its developer experience, repeatedly tripped up an AI coding agent during deployment. He transformed the tool into an AX-ready platform by adding a single non-interactive command — useless for most humans but ideal for agents.

Standards and Ecosystem Shifts

A recurring theme in Matt’s lecture was the lack of standards around agentic experience (AX). He highlighted early frameworks like the Model Context Protocol (MCP) and lightweight conventions such as llms.txt as signals of the industry’s direction.

Matt warned against naively exposing entire APIs to agents. He urged developers to treat MCP like a user interface for LLMs, carefully designing the minimal tools and context agents need to succeed.

He contrasted Salesforce’s closed approach, which pushes proprietary agents while restricting integration, with HubSpot’s open strategy, which embraces MCP to give agents seamless access. The latter, Matt argued, shows how embracing AX can differentiate products in a fast-evolving ecosystem.

For Matt, AX isn’t just a feature or protocol — it’s a discipline, like UX. It demands constant observation, iteration, and empathy, not just for human end-users but also for the agents acting on their behalf. He suggested this shift could unlock enormous economic potential by serving existing professionals and bringing entirely new demographics into web development.

At Netlify, early investments in AX have already delivered results. Since launching these efforts a year ago, the company has seen a fivefold increase in daily signups and a sevenfold jump in paid conversions. Even more importantly, AX has empowered millions of new users with little or no coding background to become active builders on the platform.

A Call to Developers

As AI agents take on more tasks — writing, debugging, and deploying code — the barriers to entry for development are collapsing. Skills like algorithm design and CLI expertise may lose weight, while qualities such as product sense, empathy, and domain knowledge gain importance.

“This is an opportunity to invite so many more people into the room,” Matt said, highlighting use cases from HR teams designing surveys to marketers prototyping campaigns with AI agents.

The next generation of developers are coming, and it’s up to us to ensure the web is ready for them.

For developers, Matt delivered a clear message: just as UX and DX once drove competitive advantage, mastering agentic experience (AX) may soon decide which tools — and which companies — define the future of software.