From Code Author to Agent Experience Designer
Agent experience design is the emerging discipline where engineers define how AI agents, systems, and humans interact so that intent expressed in natural language becomes reliable software behavior. In this new model, the primary interface is not source code but conversational intent, capabilities, and constraints that guide agents. Netlify CTO Dana Lawson describes agentic AI as the next abstraction layer, where “intent — expressed in conversational language — becomes the next programming language that exponentially more people can create with.” As AI tools take over routine implementation tasks, engineers are measured less by lines of code and more by the clarity of system boundaries, policy, and feedback loops. Code writing was already a minority of engineering work; now it becomes one of many outputs produced by agents, while the engineer curates the experience, trust, and safety of those agents across the entire software development lifecycle.
A Billion New Apps and the Software Development Evolution
AI-driven development is expected to widen the pool of builders dramatically. Lawson predicts there will be a billion new applications written by 2029 because agentic systems enable what she calls “the builder” — anyone able to express intent, not only trained developers. That wave accelerates software development evolution: planning, implementation, and testing shift into a continuous, machine-speed loop. Agents generate code, propose architectures, write tests, and open pull requests, while humans provide context and judgment. For engineers, the scarce resource is no longer bandwidth to type code but discernment about which ideas deserve to exist. Outcome Engineering describes this as moving from “what can we build?” to “what should we refuse to build?” In an environment where almost any feature can be produced quickly, the strategic edge lies in deciding where to direct agent effort and what to leave on the cutting-room floor.
Redefining the Engineer’s Role in AI-Driven Development
Lawson argues that engineers are now “the shepherd of production,” responsible for what goes in and out of complex, agent-rich systems. Instead of hand-coding each feature, they design guardrails, events, and policies that let agents act safely. Agent experience design focuses on where humans and agents collaborate across the software development lifecycle: deciding when to keep humans in the loop, what an agent is allowed to change, and how to expose system intent clearly. Netlify’s own platform rebuild offers a preview of this engineer role transformation. By treating AX as the combination of developer experience and user experience, the team clarified error messages, structured build output for machines, and removed human assumptions. The result made the system easier for both citizen developers and professional engineers, while exposing how much modern engineering now depends on curating architecture, signals, and collaboration patterns instead of typing code all day.
Architectures for Agents: Capabilities, Events, and Legibility
To support agent-native workflows, engineers must reshape core architectures. Lawson describes a shift from APIs to capabilities, where systems expose intent-level actions like create_a_site rather than low-level HTTP verbs. Alongside that, software moves from request-response patterns to event-driven streams, allowing agents to subscribe to signals, observe behavior, and act autonomously when permitted. Another change is from machine-readable to agent-legible systems: architectures must be described in ways agents and humans can both understand before changes are made. Hidden tribal knowledge in Slack threads or undocumented infrastructure modules becomes a liability. This software development evolution raises the bar for engineers: they must maintain clear capability models, reliable event taxonomies, and transparent system maps. AI-driven development works only if agents can interpret these structures; designing them becomes a central part of the job, replacing much of the manual coding once required.
Human Judgment, Guardrails, and the Future of Agent Experience
Even as agents write code and operate infrastructure, Lawson stresses that “engineering evolves from implementing every feature to really ensuring that the systems and guardrails architecture is tight and solid.” Netlify’s approach grounds agent behavior in clear human-set boundaries: each agent executes in a sandbox, actions are logged for audit and rollback, and humans remain in the loop for intent-level decisions. The aim is trustworthy AI-driven development, not fully autonomous production. Lawson warns, “If you can’t explain what the agent did, why would you trust it in production?” For engineers, agent experience design becomes the art of combining safety, performance, and environmental considerations with business goals. AI has forced teams to clarify architecture and signals, and, in Lawson’s view, “made us all better developers” by sharpening the judgment needed to build systems that agents — and people — can work with confidently.
