Artificial intelligence is changing the way people interact with technology—and it is also changing who   or what, technology is designed for.

As AI systems become capable of interpreting information, making decisions and completing tasks independently, businesses are moving beyond traditional software interfaces. Instead of relying on people to navigate every step of a process   organisations can increasingly use AI agents to carry out workflows in the background.

This evolution is giving rise to a new area of design: Agentic Experience (AX) Design.

While traditional UX focuses on creating useful and intuitive experiences for people, AX focuses on the systems and environments in which autonomous AI agents operate. It considers how these systems access information, follow rules, make decisions, handle exceptions and interact with other digital tools.

The result is a broader definition of experience design—one that considers not only what people see, but also what intelligent systems do behind the scenes.

What Is Agentic Experience Design?

Agentic Experience Design is an emerging approach to designing the environments where AI agents operate.

Traditional User Experience (UX) design is primarily concerned with human interaction. Designers create interfaces, navigation systems, forms and workflows that help people complete tasks.

AX takes a different perspective.

An AI agent may receive a goal, determine the steps needed to achieve it, retrieve information from different sources and complete actions across multiple systems. It may not need a traditional screen or interface at all.

This means designers increasingly need to consider questions such as:

• What information does an AI agent need?
• Which systems should it be able to access?
• What decisions can it make independently?
• When should it ask for human input?
• What happens when something goes wrong?
• How can its actions be monitored and evaluated?

These questions sit at the heart of AX Design.

Why AI Agents Are Changing Experience Design

For years, digital design has centred on the human user.

A typical product experience might involve a person opening an application, navigating a menu, entering information, selecting an option and completing a task.

AI agents can change this sequence completely.

Instead of completing every step manually, a person may simply provide an objective. The AI system can then determine the required actions and execute them across connected tools and platforms.

For example, an agent could potentially gather information   organise documents, update records, prepare reports or coordinate routine tasks without requiring a person to interact with every individual system.

This creates an important shift:

The experience is no longer limited to the interface a person sees. It also includes the invisible systems that allow AI to act.

The Importance of Understanding Existing Workflows

Automation is only effective when the underlying process is well understood.

Many business processes contain far more complexity than official documentation suggests. Employees often rely on experience, judgement and informal procedures to deal with unusual situations.

These details can be easy to overlook when introducing automation.

If an AI agent is given an incomplete understanding of a workflow, it may repeat mistakes quickly and consistently. A process that previously caused occasional problems could create much larger issues when performed automatically at scale.

This is why effective AX Design begins with investigation.

Designers and technology teams need to understand how work is actually performed, identify variations and exceptions and determine where human judgement remains important.

Three Key Areas of AX Design

Although the field is still developing, AX Design can be understood through three broad areas of responsibility.

Understanding the Workflow

The first step is understanding the process itself.

This involves identifying how tasks are completed, which systems are involved, what information is required and where decisions are made.

It also means looking beyond official procedures to discover workarounds, exceptions and practical challenges.

Preparing the Digital Environment

AI agents depend on access to reliable information and well-structured systems.

Data, APIs, documentation, permissions and digital platforms all need to be organised in ways that allow AI systems to interact with them effectively.

A poorly structured digital environment can limit what an AI agent is capable of doing, regardless of how advanced the underlying technology may be.

Defining Rules and Boundaries

Autonomous systems need clear boundaries.

Designers and technology teams must establish what an AI agent is allowed to do, what information it can access and when human intervention is required.

Clear safeguards can help prevent incorrect decisions, inappropriate actions and unintended consequences.

A Practical Approach to Agentic Design

Designing for AI agents requires a different starting point from conventional interface design.

Instead of immediately creating screens or prototypes, teams can begin by understanding the environment in which the agent will operate.

Map the Real Process

Document how work is actually completed rather than relying solely on existing process diagrams.

Identify people, systems, information sources, decisions and exceptions involved at every stage.

Assess Automation Opportunities

Not every process is suitable for autonomous execution.

Consider how predictable the workflow is, how frequently exceptions occur, what risks are involved and where human expertise is essential.

Establish Guardrails

Define clear conditions for success and failure.

An agent should know what to do when information is incomplete, systems conflict, unexpected situations occur or a decision falls outside its authority.

Make Complex Systems Understandable

Even when AI performs tasks in the background, people still need to understand what is happening.

Visual workflow maps, decision trees, system diagrams and monitoring tools can help teams evaluate and audit agent behaviour.

Moving Beyond Chatbots

The development of AI is often associated with chatbots and conversational interfaces, but autonomous agents represent a much broader opportunity.

Instead of simply answering questions, AI systems can increasingly interact with software, analyse information, coordinate tasks and perform actions.

This could transform areas such as customer service, administration, finance, operations, research and internal business processes.

The most important change may therefore happen behind the interface.

People may interact with AI through a simple instruction, while multiple systems work together in the background to complete the requested task.

The Expanding Role of Designers

The rise of AI agents does not mean traditional UX is disappearing.

Human-centred design will remain essential wherever people interact directly with products and services.

However, the role of designers is expanding.

Designers may increasingly need to understand data structures, system architecture, automation logic, AI behaviour and operational processes alongside traditional principles of usability and accessibility.

The focus is shifting from designing only what people see to designing the broader environment in which technology operates.

Designing the Next Generation of Digital Experiences

Agentic Experience Design represents an important evolution in the relationship between design and technology.

As AI becomes more capable of acting independently   organisations will need to think carefully about how these systems operate—not simply how they look.

Successful AI adoption will depend on more than deploying powerful models. It will require well-designed workflows, reliable information, clear rules, appropriate safeguards and meaningful human oversight.

The future of experience design is therefore likely to extend beyond screens and interfaces.

UX designs experiences for people. AX expands that thinking to the intelligent systems working alongside them.

As autonomous AI becomes a larger part of everyday digital operations, designing the invisible may become just as important as designing what users can see.