AI/UX: Experience Design of AI Agents

Summary: Welcome to a new episode of i-UX. In this month’s episode we will discuss the current lay of the AI land in terms of user adoption. Additionally, why (and how) learning to design for AI agents matters.

Industry research

Before summoning AI Agents, let’s examine the lay of the AI land based on recent research.

Despite the initial hype surrounding GenAI tools like ChatGPT, widespread adoption has been slower than anticipated. This lag can be attributed to the time required for users to learn and integrate new technologies effectively.

According to recent research from McKinsey, while AI has significantly impacted business operations, GenAI is relatively new, and its adoption necessitates the development of user-friendly applications and sufficient time for users to adapt.

Drawing parallels to the iPhone’s introduction in 2007, Kiran Prasad, McKinsey senior adviser, highlights that it took six to eight years before the iPhone became integral to users’ lives. This suggests that genAI might follow a similar adoption curve, emphasizing the importance of designing intuitive and engaging user experiences to facilitate this transition.

source: mobilelinik

Additionally, NN/g recently published an article, “AI Adoption in the Workplace Still Low, 2 Years Later” based on Pew Research. In short, it suggests that Tech Workers are more likely to use AI than non-IT workers.

These workers are more likely to work in jobs involving data processing (63%), information technology (12%), as well as industries like finance, real estate, and insurance.

Key Takeaways for Design Teams:

1. Designers should avoid assuming all users have the same AI knowledge and experience as they do (avoid being in a tech bubble and steer away from the False Consensus effect).
2. Younger workers are adopting AI more, but usability research must include all age groups and industries to avoid biased findings.
3. AI adoption is still limited, so clear onboarding, relevant prompt controls and gradual exposure (design’s graceful degradation) to new use cases are vital for engagement.

Emergence of AI Agents

This brings us to the next act of our story: AI Agents.

Enter front-stage left: AI agent (on a horse in shiny armour).

Emergence. image by author & Midjourney

It will not be long until we will be designing just for human customers. Experience Designers and Strategists will need to train to design for a new class of personas on the front-stage i.e., AI agents. Like a new species entering the ecosystem, these intelligent, autonomous systems will reshape how services are accessed, evaluated, and delivered in the real-world interfaces of B2B and B2C businesses.

AI agents in short, are intelligent systems that can figure out how to handle tasks on their own without needing constant instructions. They interact with their surroundings, collect information, and take action to reach specific goals.

source: BotsCrew

These agents “sense” what’s happening around them, using digital inputs (like questions in a chatbot). Over time, AI agents learn from past tasks, feedback from people, and other information to get better and adjust to changes (BotsCrew).

As a result, AI agents will become active participants in service ecosystems. They’ll be able to schedule meetings, interpret data, mediate communication, and increasingly negotiate and make decisions on behalf of users.

From Human-Centric to Agent-Inclusive Design

Traditionally, UX has been human-centered: optimize for clarity, emotion, ease, and delight. But AI agents won’t need lengthy microcopy or diverse aesthetic colors. What they’ll need is structured data, API access, and programmatic clarity.

It’s like designing for both tourists & GPS systems: one needs stories and signs, the other needs maps and metadata.

This shift requires designers to think about AI agents as a new persona in the design process. These agents could be personal assistants (like a user’s private AI) or organization-owned agents (like AI-powered chatbots, co-pilots, or data processors). As these systems take on more complex roles, they will shape and even filter the user experience.

Imagine two AI agents talking: one representing a user (like you), the other representing a business (like a flight centre). They compare pricing, availability, terms, and past preferences. We are not far off from this scenario. It is already being prototyped in healthcare, finance, and travel.

In this AI-mediated model, the front-stage will offer a smart negotiation between systems. The back-and-forth between human and interface could increasingly be handled by AI agents. The design challenge therefore becomes less about how something looks and more about how systems interact.

Designing the Invisible. image by author & Midjourney

Designing the Invisible: Structured, Trusted, and Clear

As this new paradigm takes shape, UX designers will need to prioritize:

• AI-to-AI compatibility: How well do systems share and interpret data?
• Trust and transparency: Can users understand and control what their AI is doing?
• Invisible UX: Interfaces may disappear but the experience will remain. We must ensure consistency, clarity, and confidence, even when machines mediate the journey.

According to Helen Armstrong, “We, designers, have to own up to our role in this emerging predictive world because we have our invisible hand meddling with the controls.” — (Big Data. Big Design)

More KPIs in the air. Image by author & Midjourney

Metrics Are Changing Too

With AI agents mediating interactions, the way we measure success will evolve too. Traditional KPIs like NPS or task-completion time won’t be enough. We will need to consider measuring:

1. Data accuracy
2. AI-to-AI task success rates
3. Trust and Transparency levels
4. User confidence in delegation

Designers must understand how these agents behave under uncertainty, how users perceive their performance, and how to design fallback experiences when things go south.

🎥 Video of the month

Introduction to the AI agent-based application tutorial

[embed]

🙃 XD humour

AI Agents

🍃 Resolution: Designing with AI, for AI, and for Humans

To conclude, UX Design and Strategy professionals should consider to have three users to serve in future: the customer, the employee, and the AI agents.

This new wave doesn’t diminish our Experience Design craft — it just deepens it.

We need to stretch our empathy, structure, and strategy skills. We need to train, build and connect systems with creative intelligence, that are not only usable and useful but also compatible with the emerging intelligence that will surround us soon. We need to prepare designing for a world where AI doesn’t just assist but also acts.

Feel free to reach out if you would like to dig deeper into elevating your team’s AI/UX skills. Stay tuned for next month’s i-UX article, where we will aim to offer more insightful Experience Design & Strategy insights and tips. Until then, stay curious and remember, There is no “I” in “UX”.

About the Author: Noman Siddiqui is an Experience Design Leader, Adjunct Professor of Design and a self-professed Usability Geek. He serves as the Experience Design & Strategy Director at Nomans Land Creative.