We need to talk about AI in UX. For some reason, the standard advice has become either “it is magic” or “it is a total waste of time,” and both perspectives are killing your productivity. I have wasted plenty of hours trying to force LLMs to do things they simply cannot do, but I have also found the sweet spot where they actually move the needle on real client projects.
If you are a business owner or a developer frustrated by slow research cycles, the bottleneck usually isn’t the thinking—it is the grunt work. After two years of daily usage, I have realized that the trick to AI in UX isn’t finding a “perfect prompt.” It is treating the AI like an enthusiastic intern with zero life experience. You wouldn’t trust them to ship code unsupervised, but they are brilliant at sorting the mail.
Why AI in UX is Like Managing an Intern
An intern has the qualifications but no context. When I work with AI, I follow a strict mental model. I define the role (e.g., “Act as a senior copywriter”), break the task into atomic steps, and define success via a rubric. Specifically, I make the AI “think” about its approach before it outputs a single word. This prevents the generic, “hallucinated” fluff that most people complain about.
One of my favorite use cases is synthesizing user research. I used to avoid open-ended survey questions because reviewing hundreds of responses was a nightmare. Now, I feed those transcripts into a project space. AI can identify recurring themes and objections in minutes. However, a major gotcha is trust—I always require the AI to cite direct quotes so I can verify the reasoning.
Speaking of research, I recently wrote about how to master robust virtual personas with this simple hack. Using AI to build “functional personas” based on actual data is a massive win compared to the old-school marketing fluff we used to rely on.
Design Audits and Predicative Visuals
AI isn’t ready to design your final UI—it still produces unpolished choices that feel “off.” But for auditing? It is lethal. Tools like Attention Insight use AI trained on thousands of eye-tracking studies to predict where users will look with 90-96% accuracy. This is a game-changer when dealing with stakeholders who insist on crowding the header with five different CTAs.
For e-commerce sites, I often use Baymard’s UX Ray. It scans your site against their database of 200,000+ hours of usability research. It doesn’t replace a human audit, but it highlights the low-hanging fruit (like checkout friction) that would otherwise take me days to document.
Automating the UX “Small Stuff”
As a developer, I use AI for the low-risk “throwaway” tools that I used to spend hours refactoring. For example, a client recently needed their blog posts to show a “Last Evaluated” date automatically to maintain trust. Instead of hunting for a plugin with bloat, I had AI help me draft a simple, lightweight function.
<?php
/**
* Automatically append 'Last Evaluated' date to UX articles.
* Prefixing with bbioon_ for safety.
*/
function bbioon_append_last_evaluated_date( $content ) {
if ( is_single() && in_the_loop() && is_main_query() ) {
$updated_date = get_the_modified_date();
$evaluated_html = '<p class="ux-meta">Last UX Review: ' . esc_html( $updated_date ) . '</p>';
return $evaluated_html . $content;
}
return $content;
}
add_filter( 'the_content', 'bbioon_append_last_evaluated_date' );
Is this groundbreaking? No. But it took 30 seconds to ship it. Furthermore, AI lets me build ROI calculators and scripts for top task analysis that would have been “nice-to-haves” in the past because of budget constraints.
Fixing the “Terrible Copy” Bottleneck
The biggest experience-killer on any WordPress site is bad copy. I have stopped asking clients to write “web copy.” Instead, I ask them for bulleted answers to specific user questions. I then use AI to draft the content based on a strict readability guide—usually a ninth-grade level using the Hemingway Editor logic.
This workflow removes the “blank page” fear for stakeholders and ensures the site follows proven UX research patterns that actually convert. Consequently, the project moves faster, and I don’t have to spend my time fixing awkward jargon.
Look, if this AI in UX stuff is eating up your dev hours, let me handle it. I’ve been wrestling with WordPress since the 4.x days and know exactly where these tools help and where they fail.
The 30% Speed Increase
AI has likely given me a 25% to 33% increase in output. A report that used to take five days now takes a few hours. That doesn’t mean I am working less; it means I am spending more time on strategic thinking and less on manually reviewing survey transcripts.
Your value as a professional lies in the questions you ask and the creative leaps you make. AI cannot innovate; it can only recognize patterns. Start small, treat it like that enthusiastic intern, and focus on the problems that actually require your human brain to solve.
Leave a Reply