AI Is Your New Intern, Not Your Substitute

Half 1 of the “UX × AI” collection.

Foreword: Why this collection exists, and why I’m writing it

Earlier than I make any argument, let me inform you one thing private.

I’ve been on this discipline for over 25 years. I’ve watched UX take up wave after wave of disruption — the shift from desktop to net, from net to cellular, from cellular to voice, and now from voice to AI. And in each wave, I’ve noticed the similar sample repeat itself with exceptional consistency.

A strong know-how arrives. The business splits into two camps. One camp broadcasts that every part will change perpetually and that those that resist are completed. The opposite camp dismisses the know-how as overhyped, defends the establishment, and waits for the noise to subside. And someplace in the center — the place the place trustworthy, grounded, practitioner-level pondering ought to reside — there is nearly nothing.

That hole is what this collection is constructed to fill.

I’m not writing “UX × AI” as a result of I’m enthusiastic about AI as a know-how. I’m penning this as a result of I’m deeply involved about how the design neighborhood is responding to it — with both uncritical enthusiasm or defensive worry — and since I imagine that design professionals deserve higher than both possibility. They deserve a collection that thinks clearly. That acknowledges the real uncertainty. That gives frameworks for making choices in that uncertainty. And that is trustworthy sufficient to say what it does not know alongside what it does.

Over 25 years, I’ve constructed UXExpert, UXUITrainingLab, UXTalks, LucyUX, and UXUI Summit. I’ve labored with designers, researchers, product managers, and design leaders throughout India and internationally. I’ve mentored a whole bunch of UX professionals at each profession stage. And the query I’m being requested greater than every other proper now — by skilled designers, by early-career researchers, by design leaders navigating crew technique — is some model of the similar query.

What does AI imply for me? For my work? For my worth? For my profession?

This collection is my reply. Not a reassuring reply. Not a daunting one. An trustworthy one.

What this collection will cowl, and the way

“UX × AI” runs throughout ten articles, structured round 4 editorial pillars that rotate and mix throughout the collection.

• Studying Curve covers the foundational ideas — the psychological fashions and vocabulary that make every part else on this collection make sense. What AI really is, in the UX context. How to take into consideration prompting as a design talent. The distinction between augmentation and automation. The constructing blocks {that a} practitioner wants before they’ll consider anything intelligently.
• Delusion Busters is the place I might be most direct. The design neighborhood is at present working below a set of narratives about AI that are distorting choices at each stage — particular person, crew, and organizational. These narratives want to be examined, examined in opposition to proof, and, the place obligatory, dismantled. Delusion Busters does that work, with reasoning slightly than assertion.
• How to Undertake AI interprets understanding into follow. Not theoretical frameworks for AI integration — sensible, workflow-level steerage for the designer, the researcher, the design chief who is prepared to transfer from curiosity to competence. What to do first. How to consider instruments. How to construct fluency with out overhauling your total follow directly.
• At this time and Future zooms out. What is the AI instruments panorama really wanting like proper now, and the way ought to practitioners give it some thought? What is agentic UX, and why does it matter? What are the moral {and professional} stakes — in authorship, in possession, in the way forward for the self-discipline — that the business is not but taking severely sufficient?

This first article belongs to the first two pillars concurrently. It is foundational — it establishes the psychological mannequin that the remainder of the collection builds on. And it is a myth-buster — as a result of the fantasy it addresses is the one inflicting extra hurt to the design neighborhood than every other proper now.

The parable that AI is coming to change you.

“The query is not whether or not AI will change UX. It is going to. The query is whether or not UX professionals will form that change or be formed by it.” — Tushar Deshmukh, UXExpert

The worry is actual. The body is improper

Let me start by honoring one thing before I argue in opposition to it.

The worry is actual. I need to say that clearly and with out qualification — as a result of dismissing it will be dishonest, and honesty is the dedication this collection is constructed on.

A UX designer in Bengaluru who has spent eight years constructing her portfolio, her analysis abilities, and her skilled popularity is not being irrational when she seems at the tempo of AI improvement and feels her chest tighten. A mid-career UX researcher in Pune who has constructed his total skilled identification round the craft of qualitative analysis is not being dramatic when he wonders whether or not AI will make that craft redundant. A design chief in Mumbai who is being requested by her group to justify headcount in a world the place AI can generate wireframes in seconds is not imagining the strain she is below.

The worry is smart given the narrative that surrounds it. And the narrative is nearly totally improper.

The dominant story — AI will change UX designers — is constructed on a class error so elementary that it leads each dialog it enters in the improper path. The error is this: it conflates duties with roles. It identifies a listing of issues that designers do — generate wireframes, write microcopy, create person flows, and synthesize analysis notes — notices that AI can carry out variations of these duties, and concludes that the function of the designer is subsequently below risk.

This is like watching a surgical robotic make incisions extra exactly than human fingers and concluding that surgeons are out of date. The incision is a activity. The surgeon’s function — the medical judgment, the analysis in the ambiguous case, the choice when the operation reveals one thing surprising, the dialog with the affected person’s household after — is not a activity. It is a constellation of experience, contextual intelligence, moral accountability, and human judgment that the robotic does not have and can’t have.

The identical is true of UX design. The wireframe is a activity. The design function — the person analysis that reveals the downside price fixing, the methods pondering that understands how a design choice in a single a part of a product creates penalties in one other, the facilitation that builds stakeholder alignment round a user-centered choice, and the judgment that acknowledges when an AI-generated answer seems proper however is improper for this particular context and this particular person — that is not a activity. It is an expert competence. And it is not going away.

What is altering is the work floor on which that competence operates. And that is a really completely different factor.

Meet your intern

I need to provide a reframe. One which I’ve discovered genuinely helpful — not as a comforting metaphor however as a virtually correct description of what AI instruments really are, in the context {of professional} UX follow.

Consider AI as your new intern.

Not a superhuman intern. Not a threatening intern who secretly is aware of greater than you and is angling in your place. An actual intern — one who brings particular spectacular capabilities, important actual limitations, and a whole dependence on you for path, analysis, correction, and accountability.

Your intern is quick. Genuinely, impressively quick at sure sorts of technology. Ask them to produce ten variations of a button label, and they’re going to have them in seconds. Ask them to draft a first-pass sitemap construction for a given content material kind, and it’ll seem before you’ve got completed the temporary. Ask them to pull aggressive examples of onboarding patterns, and they’re going to synthesise extra in two minutes than you might collect in two hours of guide analysis. This velocity is actual, and it is priceless.

Your intern is tireless. They do not really feel the artistic fatigue that units in when a human designer has been observing the similar downside for 4 hours. They do not carry emotional funding in yesterday’s iteration, which makes it onerous to abandon it at present. You possibly can ask them to attempt one thing utterly completely different — a special tone, a special construction, a special design path — and they’re going to try it with out the resistance that comes from having labored on the earlier model.

Your intern has learn every part. They’ve been skilled on extra design documentation, analysis literature, UX case research, accessibility tips, design system specs, and interplay sample libraries than any human being might learn in 100 lifetimes. Once you ask them about established patterns for a given use case or about the accessibility implications of a design choice, they draw on a information base of extraordinary breadth.

And your intern wants fixed supervision. At all times. With out exception.

Your intern has by no means sat throughout from a person. They’ve examine customers — hundreds of thousands of phrases about customers, from analysis stories, ethnographic research, and usefulness check transcripts. However they’ve by no means been in the room when a 72-year-old man in a rural district of Rajasthan tries to navigate a authorities portal on a battered smartphone with a cracked display screen and an information connection that cuts out each three minutes. They’ve by no means watched a first-generation school pupil in Nagpur abandon a banking app not as a result of they do not perceive it — they do — however as a result of the language of the interface communicates, with out stating it, that it was not designed for somebody like them. They’ve by no means felt the factor that expert UX researchers really feel when an interview takes an surprising flip and divulges one thing true a couple of human want that no temporary, no persona, and no dataset would ever floor.

Your intern does not perceive context. They’ll generate design patterns for onboarding flows. They can’t know that this particular onboarding stream is for a product used primarily by healthcare employees in Tier 3 cities, who’ve deeply particular privateness issues formed by the skilled sensitivity of the information they are coming into, issues that your intern has processed as information factors however can not perceive as lived expertise. Context — actual, particular, located, human context — is what the designer carries into each venture. The intern does not carry it. They obtain it from you in the type of a immediate. And a immediate is a big discount of context, not a full switch of it.

Your intern can’t be accountable. When the design ships and one thing goes improper — when the stream that seemed appropriate in Figma fails in the fingers of actual customers in a approach that causes real hurt — the intern is not in the room. The designer is. The skilled and moral weight of the work belongs to the individual whose judgment formed it. And that individual is you.

“The best design groups are not selecting between people or AI. They are creating collaborative workflows that use AI to increase the designer’s capabilities — not to change the designer’s judgment.” — Tailored from Optimum Workshop, 2025

What the analysis really tells us

The nervousness about AI and design jobs has been amplified by a media atmosphere that rewards dramatic predictions over cautious evaluation. The precise analysis, when examined with out the drama, tells a extra particular and extra helpful story.

Nielsen Norman Group’s 2025 UX Reset report — one in every of the most rigorous assessments of AI’s affect on the UX career accessible — does not conclude that UX professionals are being changed. It concludes that the bar for what makes a UX skilled indispensable is rising. The instruments that deal with repeatable, execution-heavy duties are turning into extra succesful, which creates strain on practitioners who spend most of their time on these duties. However practitioners whose worth was at all times rooted in higher-order capabilities — strategic analysis design, cross-functional management, and design judgment that can’t be lowered to a sample — are not below risk. If something, as Nielsen Norman Group’s 2026 State of UX report argues, the crucial is to design deeper, not to design quicker. The depth of perception, the high quality of empathy, the rigor of methods pondering — these change into the differentiators exactly as a result of AI has commoditized the surface-level outputs.

Optimum Workshop’s 2025 analysis on AI in UX follow discovered that AI is eliminating the parts of analysis work that researchers discover most tedious and least intellectually participating — transcription, preliminary sample coding, and large-scale survey synthesis. This is not a alternative. This is reallocation — of human consideration away from mechanical processing and towards the interpretive, relational, and strategic work that requires human intelligence. The researcher who is freed from three hours of transcription does not have three fewer hours of priceless work. They’ve three extra hours to do the work that solely they’ll do.

McKinsey’s analysis on human-AI collaborative groups discovered that groups integrating AI analysis instruments spend considerably extra time on strategic planning and synthesis and considerably much less time on execution. Once more — not a alternative. Reallocation. The human is spending extra of their time on the work that people do greatest.

The UX job market contracted sharply in 2024. However as cautious evaluation from ROSSUL and others has established, the main trigger was not AI displacement — it was a post-pandemic financial correction, compounded by know-how sector-specific restructuring, with AI adopted as a handy explanatory narrative by organizations making cost-cutting choices that had monetary slightly than technological roots. The excellence is not merely semantic. An financial cycle is momentary. A structural technological displacement is everlasting. And the proof factors clearly towards the former.

The talents that AI makes extra priceless, not much less

Right here is the perception I need each designer, researcher, and design chief studying this to carry ahead. It is the most virtually essential factor on this article.

AI does not devalue human abilities uniformly. It devalues a selected class of abilities — these closest to sample technology, template manufacturing, and repeatable execution. And it dramatically will increase the worth of a special set of abilities — the ones which have at all times been the deepest a part of UX follow, however which have generally been troublesome to articulate in a world the place deliverable manufacturing was the most seen output.

• Real empathy turns into extra priceless. Not empathy as a phrase in a job description. Empathy is the disciplined follow of understanding one other human being’s expertise with enough depth and specificity to make design choices on their behalf. AI processes information about customers. It synthesises patterns from massive datasets about customers. Empathy means sitting with a selected individual in a selected second and being modified by what you find out about their expertise. That is not an information processing activity. It is a human one.
• Analysis design turns into extra priceless. AI can analyze analysis information on a big scale. It can not design a analysis research that generates information price analyzing. The craft of realizing which methodology will really reply the query you want answered — how to assemble a dialogue information that surfaces trustworthy responses slightly than socially fascinating ones, how to recruit members who genuinely characterize the inhabitants you are designing for, and the way to acknowledge when a analysis discovering is actual versus when it is an artifact of your methodology — is experience that deepens with deliberate follow and can’t be generated from a immediate.
• Methods pondering turns into extra priceless. The power to see not simply how a single display screen capabilities however how a whole product ecosystem behaves — how a design choice in a single a part of the system creates penalties in one other, how person journeys span channels and contexts and time, how at present’s design choice shapes tomorrow’s technical constraint — this is the competence that distinguishes the senior UX practitioner. AI can generate parts. It can not architect a system with the knowledge of somebody who has watched methods fail and understood why.
• Facilitation and organizational affect change into extra priceless. Design does not reside in Figma. It lives in the assembly the place the product choice is really made. The power to deliver stakeholders to real alignment on a user-centered path, to make a compelling case for design high quality below industrial strain, and to navigate the political and cultural dynamics that decide whether or not a great design ships or will get overridden — these are deeply human capabilities. They are capabilities that the designers who’ve invested in them, who’ve constructed their careers on them, will discover more and more central as AI handles extra of the execution.
• Moral judgment turns into extra priceless. As AI is built-in into design follow, the questions of what needs to be designed — not simply what may be designed — change into extra pressing. The designer who understands accessibility not as a compliance requirement however as a justice dedication. The researcher who acknowledges when an information assortment follow raises moral issues that the authorized crew has not thought of. The design chief who asks whether or not an AI-generated suggestion serves the person or merely optimizes the metric. These are judgment capabilities. They can’t be automated. And their worth solely will increase as the methods round them change into extra highly effective.

The three errors I see designers making proper now

In my work with designers throughout India and internationally — in mentoring periods, in workshops, in the conversations that occur when the leaders are not in the room — I see three particular errors being made in response to AI. I need to title them immediately, as a result of they are inflicting extra hurt than the disruption that individuals are afraid of.

• The primary mistake is ready. A major variety of designers are responding to AI by hoping it can stabilize before it requires them to change. By persevering with to work precisely as they’ve, watching the developments from a distance, and assuming they may have time to adapt when the panorama settles. It is going to not settle in a kind that permits this technique to work. AI is not a development to be waited out. It is infrastructure — like the web, like cellular — that is being constructed into the basis of each design device, each product improvement course of, and each shopper expectation. The designers who are constructing real fluency now — who are experimenting with AI of their workflows, who are creating a practitioner-level understanding of what these instruments do and do not do — might be measurably more practical when AI fluency turns into a baseline expectation. That second is nearer than the ready technique assumes.
• The second mistake is outsourcing judgment. At the reverse excessive, some designers have embraced AI so utterly that they are utilizing it to make choices that ought to not be delegated. Accepting AI-generated person personas with out validation from actual customers. Transport AI-drafted copy with out asking whether or not it really communicates with the heat and precision the person deserves. Approving AI-produced design patterns with out evaluating whether or not they are acceptable for the particular cultural, accessibility, and contextual necessities of this product and this person. AI is a strong technology device. It is not a judgment device. Once you outsource judgment, you outsource accountability — and the accountability belongs to you.
• The third mistake is letting worry form the dialog. In management conferences, in crew discussions, in conversations with shoppers and stakeholders, the worry narrative is distorting choices. Design leaders are making AI adoption decisions primarily based on nervousness slightly than evaluation. Designers are framing their skilled worth defensively — arguing in opposition to alternative slightly than articulating what they bring about. The dialog wants to shift. Not from “AI received’t change us” — a defensive body that assumes the risk is actual and argues in opposition to it. To “Right here is precisely how we are integrating AI to do higher work for our customers and our organizations” — a assured, forward-looking body that positions AI as a functionality multiplier slightly than an existential risk.

What glorious human-AI design collaboration seems like

Some practitioners and groups are getting this proper — not by both resisting AI or deferring to it, however by creating a clear-eyed, competence-based relationship with it that amplifies their human capabilities.

The UX researchers who use AI transcription and preliminary coding as a place to begin, then apply their knowledgeable judgment to interpret the patterns the AI has recognized, problem the interpretations that really feel reductive, and floor the insights that the AI’s pattern-matching can not attain. These researchers are producing extra analysis in much less time and higher analysis — as a result of the time they’ve saved on mechanical processing is reinvested in deeper interpretation.

The interplay designers who use AI to generate a broad discipline of design choices shortly — ten instructions slightly than two, fifty variations slightly than 5 — after which apply their design judgment to choose, mix, and refine from that discipline. These designers are utilizing AI to increase the answer area they discover, whereas protecting their very own craft and judgment at the heart of the choice and refinement course of.

The design leaders who are integrating AI into their crew’s workflow by mapping their present processes, figuring out the place AI can speed up execution with out compromising judgment, and investing the time saved in the strategic and relational work that AI can not do — stakeholder schooling, cross-functional collaboration, and the sluggish and obligatory work of constructing a design tradition inside a corporation that does not but have one.

What these practitioners share is not a selected device or a selected workflow. It is a transparent sense of what they bring about that AI does not — and a real dedication to utilizing AI to free extra of their time and a spotlight for the work that solely they’ll do.

Making use of LucyUX to the AI query

The LucyUX framework — Pay attention, Perceive, Conceptualize, Yield — which I’ve developed over many years of follow and utilized in the Molecular Biology and Agriculture collection, takes on a selected form when utilized to the query of how design professionals ought to relate to AI.

• “Pay attention” means listening to what AI instruments are really telling you whenever you use them — not what the advertising materials claims they do, however what they really produce. The place does the output shock you with its high quality? The place does it miss in ways in which reveal its elementary limitations? The place does it produce one thing that appears proper however feels improper — and what is the supply of that feeling? The practitioner who listens rigorously to AI output, slightly than accepting or rejecting it wholesale, develops a selected and helpful understanding of what these instruments are really good for.
• “Perceive” means constructing an correct psychological mannequin of how AI methods work — not a technical mannequin, however a sensible one. Understanding that AI generates outputs by pattern-matching in opposition to coaching information, not by reasoning from first ideas. Understanding that it has no entry to the particular context of your person, your group, your product, and your design downside. Understanding that its confidence in an output has no relationship to the output’s correctness — that AI may be improper with the similar tone it makes use of when it is proper. This understanding is the basis of utilizing AI critically slightly than deferentially.
• “Conceptualize” means designing your AI-integrated workflow intentionally — not adopting AI instruments as a result of others are utilizing them, however pondering via which particular duties in your particular follow would genuinely profit from AI acceleration, which duties require human judgment that AI can not approximate, and the way the handoffs between AI-generated and human-evaluated work needs to be structured. The workflow design is itself a design downside. Apply design pondering to it.
• “Yield” in the AI context is measured in the high quality of outcomes for customers — not in the effectivity of the design course of. Did the AI-accelerated analysis course of produce insights that formed higher design choices? Did the AI-assisted ideation course of open design instructions that the crew would not have explored with out it? Did the AI-augmented workflow create capability for the human work — the deep person understanding, the methods pondering, and the organizational affect — that produces genuinely higher person experiences? These are the yields that matter. Effectivity with out high quality enchancment is not a yield. It is a trade-off which will not be price making.

“Expertise is greatest when it brings individuals collectively.” — Matt Mullenweg

When AI is handled as the designer, everybody loses

The price of misusing AI in design follow is not hypothetical. It is starting to be documented — in merchandise which have shipped with AI-generated interfaces that exclude important person populations, in analysis findings that AI synthesis has flattened into generalities, and in design choices that AI suggestion has optimized for metrics at the expense of real person wellbeing.

The designer who makes use of AI to generate a persona after which designs for that persona with out person analysis has not saved time. They’ve produced work that is doubtlessly worse than if that they had not used AI in any respect, as a result of the AI persona carries the assured presentation of knowledge with out the substance of precise human remark. The boldness is a lure.

The group that adopts AI as a design perform — changing UX analysis with AI-generated insights, changing design judgment with AI-generated patterns — has not change into extra environment friendly. It has misplaced the functionality that generates the most worth in design: the real, particular, contextually located understanding of actual customers that solely human analysis can produce.

The design chief who justifies headcount discount on the foundation that AI can deal with the work is having a bet that the duties AI has automated had been the duties that created worth. In lots of circumstances, they had been not the duties that created worth. The worth was in the judgment, the perception, the facilitation, the affect — the human work. And the human work now has nobody to do it.

Your motion this week

Understanding this reframe is helpful. Placing it into follow is extra helpful. Right here is a concrete place to begin — not a complete AI adoption technique, however a primary step that can educate you one thing actual about the precise relationship between your follow and these instruments.

Take one activity from your present workflow — one that you simply carry out often, and that entails a big technology part. Analysis synthesis. Microcopy drafting. Aggressive evaluation. Person stream mapping. Run it alongside an AI device this week. Not as an alternative of your regular course of — alongside it.

Then do the important comparability. The place did the AI produce one thing that saved you real time with out sacrificing high quality? The place did its output miss one thing that you simply knew from skilled expertise — and what is the nature of that information, the information that the AI did not have? The place did it produce one thing that seemed appropriate however felt improper — and the way did you establish that wrongness? What was the evaluative course of?

That comparability — the act of holding AI output in opposition to your personal skilled judgment and understanding exactly the place and why it falls brief — is the core competency. Not prompting talent. Not device choice. The power to know what attractiveness like in your particular context and to use AI as a quick generator that you simply apply that judgment to. That is the competence that makes you more practical and extra indispensable, not much less.

My perspective: What I really imagine

I need to shut this primary article the approach this collection is dedicated to closing each article — with a direct assertion of perspective. Not a diplomatic hedge. Not a “on one hand, on the different hand.” A place.

AI is the most vital device that has entered UX follow in the previous twenty years. It is genuinely highly effective. It is going to genuinely change how design work is accomplished. And it is not — not now, not in the foreseeable future — a alternative for the human practitioner whose judgment, empathy, contextual intelligence, and moral accountability are the precise supply of worth in design.

The designers who will thrive in the subsequent decade are not the ones who resist AI nor the ones who defer to it. They are the ones who develop a clear-eyed, practitioner-level understanding of what these instruments do properly, what they do not do, and the way to combine them right into a follow that continues to be anchored in the factor that AI can not replicate: real look after the human being on the different aspect of the interface.

That has at all times been what UX is about. A strong new device does not change what UX is about. It modifications the circumstances below which UX does its work. The circumstances are altering quicker than many people would love. The work stays the similar.

Your intern is quick. You are the designer.

Act prefer it.

Up subsequent in the “UX × AI” collection: “The Immediate Is the New Transient.” Designers already know the way to articulate intent with precision, constraint, and inventive path. They do it each time they write a design temporary, a analysis temporary, or a artistic path doc. It seems that prompting an AI is a talent that appears remarkably like this, and designers are higher positioned to develop it than nearly anybody else in the constructing. In the subsequent article, we discover why prompting is a design talent, how to develop it, and what it means for the relationship between design professionals and the AI instruments that are turning into a everlasting a part of the workflow.

References & additional studying

• The UX Reckoning: Prepare for 2025 and Beyond, Nielsen Norman Group.
• State of UX 2026: Design Deeper to Differentiate, Nielsen Norman Group.
• Using AI for UX Work: Study Guide, Nielsen Norman Group.
• A Research Agenda for Generative AI in UX, Nielsen Norman Group.
• How AI Is Augmenting, Not Replacing, UX Researchers, Optimum Workshop.
• The Rise of the Human–AI Workforce, McKinsey & Firm.
• Superagency in the Workplace: AI Report 2025, McKinsey & Firm.
• How AI Is Changing What It Means to Be a UX Designer, ROSSUL.
• Will AI Replace UX Designers? 2025 Update, Looppanel.
• Will AI Replace UX Designers?, UX Design Institute.
• How I Use Generative AI for UX Research in 2025, UX Collective.
• The Design of Everyday Things, Don Norman.
• Usability Heuristics for User Interface Design, Jakob Nielsen.
• 100 Things Every Designer Needs to Know About People, Susan Weinschenk.
• LucyUX Process: Listen, Understand, Conceptualize, Yield, Tushar Deshmukh.

Featured picture courtesy: Roman Budnikov.