Building AI Products? Why UX Is Essential

As AI projects gain traction and SaaS ventures become less popular to start, I’ve noticed a new mindset emerging-the belief that you can build an entire AI project on your own. The idea is that you don’t need a designer or even analysts to help shape your groundbreaking AI product because there’s now an abundance of AI tools, like Midjourney and DALL·E, that can assist with the process. But let me tell you-you couldn’t be more wrong, and I’m about to prove it.

So, why do you need a designer in the first place in this new era of AI tools for your AI project? I’m going to break it down and cover the most important aspects-focusing on the key principles behind creating truly great AI products.

AI Products Have More Functionalities Than Just Chat

Most AI products come with additional features that determine whether they are truly usable or not. Take image generation tools, for example-they offer a wide range of options like vary subtle, vary strong, or even a seed ID. But let’s be honest-does anyone even know what that is, anyway? ChatGPT is fairly simple since it’s just text-based, but when you move into text-to-image or text-to-video interfaces, it quickly becomes clear that a basic text field isn’t enough. Sure, you can use it by just typing in prompts, but without proper use of functionalities, you won’t be able to tap into its full potential.

Another example-take the Share feature in ChatGPT. Let’s face it, it just doesn’t work! Whether you’re trying to share a conversation with someone else or even a teammate, it’s always a hassle. More often than not, you’ll end up copying and pasting the answer instead because OpenAI clearly didn’t test or design this feature with usability in mind or is it just strategy? (evil laugh, not mine of course). To sum up, AI products are usually based on Writing functionalities, but not fully. You will still need to design a proper user flow and user interface with a UX designer, before starting the development of a feature.

‍

Midjourney settings interface showing options for image size, aesthetics (stylization, weirdness, variety), model version selection, and speed mode, displayed on a clean white dashboard layout — Midjourney user interface

‍

Pricing segmentation and upsell

If you’re building a groundbreaking AI product, you’ll definitely want to monetize it somehow (those API credits can be a real hassle to pay for, right?). This means you’ll need a solid market strategy, a well-designed pricing structure, and upsell functionalities that encourage users to move up tier-not down.

This is where a skilled UX designer becomes invaluable. Figuring out the best places to introduce upsells and designing a smooth, effective funnel isn’t just an afterthought-it’s a crucial part of your product’s success. Because let’s be real, aside from great AI functionality, your second biggest priority is making sure users actually want to pay for it.

‍

A Midjourney upgrade pop-up showing pricing and premium features, including no daily limits, artistic control, commercial license, private images, fast generation, creative upscale, and unlimited uploads. A large abstract illustration appears on the left side, with upgrade pricing starting at $10 per month on the right — In order to generate more images in Recraft at once you need to upgrade

‍

Great UI doesn’t automatically translate to great UX (even with AI generated graphics)

There are countless tools available for creating stunning assets-take Midjourney, for example, which we even shared tips and tricks for in one of our articles. Yes, you can use AI tools like this to generate benchmarks or even the visual assets you need for your project with ease.

But remember-UI is not UX. A visually impressive interface won’t save a poorly designed experience. If the entire user journey isn’t thoughtfully crafted, with all edge cases considered, you’ll likely run into serious user retention issues down the line. A beautiful UI means nothing if users get frustrated and leave.
‍

A dark-themed AI interface showing a welcome screen with a colorful 3D brain illustration on the right. The left side includes the text ‘Welcome to AI,’ a prompt input field, and a blue ‘Generate’ button. Below are three feature cards labeled Chat, Images, and Documents. — Super quick presentation of ChatGPT "generate a beautiful UI for an AI product" output, but what is the product about?

‍

Introducing AI features for Fintech and Health care

Some industries require extreme precision in user input and output, meaning they might not benefit from AI-driven features as much as others. Imagine you're planning a detailed travel itinerary for clients in a city you know nothing about. You could spend hours researching, or you could use an AI writing tool to generate ideas you might never have thought of on your own. In this case, AI is a helpful assistant.

Now, imagine using the same AI-generated interface to initiate a bank transaction or transfer money to another person or institution. The potential for errors is enormous-and the consequences could be serious. That’s why industries that rely on precision, like healthcare and fintech, will integrate AI carefully and sparingly. Many of the interfaces we use today in these fields will likely remain unchanged, even after the AI revolution.

If you're building an AI product in fintech or healthcare, it’s critical to carefully plan functionalities with a dedicated UX designer-ideally, one who specializes in the industry. At UX GIRL, fintech is one of our core specialties. We’ve honed our skills through years of experience and have mastered the craft of creating highly usable fintech interfaces.

‍

Slide titled ‘Use cases for AI in fintech’ listing examples including credit risk assessment, fraud detection, virtual assistants, AI-based personal finance tools, and algorithmic trading and portfolio management — IBM Research proposes following use cases for fintech to use AI

‍

Focus on Your Use-Who Is Still Human

The tools might be evolving, but remember-you’re still designing for human beings, and not everyone effortlessly navigates the ever-changing world of new tech. Sometimes, it’s better to stick with familiar interaction patterns rather than reinvent the wheel.

Take Windows ME, for example. Microsoft tried to revolutionize the operating system world, but instead of being embraced, it became one of the most infamous failures in tech history. Users found it confusing, buggy, and unintuitive, leading many to downgrade to older versions or simply wait for a completely new release rather than adapt to it. The lesson? No matter how advanced your AI product is, if the user experience isn’t intuitive and user-friendly, people won’t hesitate to abandon it.

So maybe it’s best to hire another human to design for a human-not a machine.

‍

A man working on a laptop in a bright, modern office filled with plants. Another person works at a desk in the background, and a large monitor displaying design software is visible on the right

‍

Afterword: The AI Revolution Needs Great UX-And So Do You

AI is changing the way we build products, but one thing remains the same-your users are still human. No matter how advanced your AI tool is, if people can’t navigate it easily, trust it, or see its full potential, they won’t use it.

That’s why great UX design isn’t optional-it’s essential. From crafting seamless user journeys to designing intuitive interactions, a skilled UX designer ensures that your AI product isn’t just powerful but also usable and profitable.

At UX GIRL, we specialize in creating AI-friendly, user-first designs, particularly in high-stakes industries like fintech and healthcare. Whether you're building a cutting-edge AI tool or enhancing an existing product, we can help you design experiences that delight users and drive business success.

Need a UX designer for your AI product? Let’s talk. Get in touch with UX GIRL today and let’s build something great together!

Artificial intelligence is no longer a distant promise for design teams-it is already reshaping how designers think, collaborate, and create. This shift was the focus of a talk delivered by Magdalena Ostoja‑Chyżyńska, CEO & Founder of UX GIRL, during Data Science Summit, one of the key events bringing together experts from data, technology, and digital innovation.

In her presentation, “How AI and Enhanced Data Access are Transforming Today’s Design,” Magdalena explored how artificial intelligence is influencing modern design practice-not as a replacement for human creativity, but as a force that is redefining how design teams work with data, insights, and complex business requirements

Two women standing at a conference venue in front of large illuminated ‘#DTS’ letters in green and purple lighting, wearing event badges and smiling at the camera

Design at the Intersection of AI and Data

The talk addressed a challenge many organizations currently face: how to integrate AI into design processes without reducing originality or oversimplifying complex user problems. As Magdalena explained, the growing accessibility of data and AI models has fundamentally changed how designers approach tasks such as briefing, user research, insight synthesis, requirements definition, and asset creation.

Rather than treating AI as a purely visual or generative tool, the presentation positioned it as a broader design accelerator-one that influences decision-making long before the first interface is drawn.

Insights from Real Client Projects

A key strength of the session was its grounding in real business practice. Drawing from ongoing client work at UX GIRL, Magdalena shared observations from testing different AI tools and models across multiple stages of the design process. These experiments focused on understanding where AI genuinely supports creative and analytical work, and where its limitations become visible in real-world conditions.

During the talk, she referenced commonly used tools such as Midjourney, ChatGPT, Claude, and Recraft, explaining how they were evaluated not in isolation, but in combination with different types of data and project constraints. The emphasis was not on novelty, but on effectiveness-how these tools behave when confronted with incomplete data, ambiguous requirements, or complex stakeholder expectations.

Creativity, Control, and the Role of Data

One of the central themes of the presentation was the relationship between AI output and data quality. Magdalena highlighted that AI-driven design outcomes are only as strong as the data and context provided to the models. Enhanced access to data can dramatically improve speed and clarity, but it also increases the responsibility of design teams to curate, interpret, and challenge that data rather than accept AI-generated results at face value.

The session made it clear that AI does not remove the need for designers’ judgment. On the contrary, it amplifies the importance of critical thinking, domain knowledge, and ethical responsibility in design decisions.

Why This Talk Resonated at Data Science Summit

Presenting this topic at a data-focused conference was intentional. The session connected two worlds that often operate separately: design and data science. By showing how AI is already embedded in everyday design workflows, Magdalena demonstrated that design maturity today increasingly depends on data literacy and cross‑disciplinary collaboration.

For many attendees, the talk offered a rare perspective-AI discussed not from a purely technical standpoint, but through the lens of practical design leadership and real client constraints.

Looking Ahead

The presentation reinforced UX GIRL’s position at the intersection of design, data, and emerging technology. Rather than following trends, the studio actively tests and evaluates new tools in live projects, translating experimentation into informed design decisions.

As AI continues to evolve, the questions raised during this session remain highly relevant: how to preserve originality, how to use data responsibly, and how to ensure that technology strengthens-not flattens-the impact of design.

For those interested in how AI is shaping the future of design beyond surface-level automation, the work and insights shared by UX GIRL offer a grounded and experience-driven perspective.

Software development is entering a new era. Writing code line by line? That’s old school. Today, you vibe code - flowing quickly with AI-assisted tools, generating features in minutes, and iterating with lightning speed.

But here’s the catch: without good design, vibe coding falls apart. No matter how fast you code, if your UX and UI aren’t rock-solid, your product will hit friction fast.

That’s why we at UX GIRL created our new service: Design for Vibe Coding. Because now that AI can write your code, great design is your real competitive edge.

A split-screen comparison infographic in a beige, pink, and black color palette. The left side, "Vibe Coding Without Design," shows a dark, chaotic scene where an AI robot generates a mess of tangled wires and broken interfaces, labeled with "Messy Code" and "Broken UX," leading to a frustrated user. The right side, "Vibe Coding + UX GIRL Design," shows a clean, organized scene where the AI’s work is channeled through a "UX GIRL" funnel into modular blocks and structured blueprints, resulting in a happy user and a "Superpowered Launch."

Vibe Coding Isn’t Just Fast Code - It’s a New Way to Build

Vibe coding is the rising mindset in modern product teams - a way of working that’s fast, fluid, and creative. It's enabled by AI tools like GitHub Copilot, Replit Ghostwriter, and Codeium, which make coding feel more like jamming than engineering.

But here's the truth: AI can help you write code, but it can’t fix a broken UX. Without the right flows, component structure, and interaction logic, your fast code becomes messy code - and the vibe is gone.

Design Is Now the Foundation of Speed

According to McKinsey, companies that prioritize design outperform their competitors by up to 32% in revenue and 56% in total returns to shareholders . Forrester also reports that every $1 invested in UX brings up to $100 ROI.

In other words: code is cheap, but design drives results.

When your product is built on solid UX and clean UI, vibe coding becomes a superpower. You eliminate friction, cut dev time, and accelerate iteration - all without losing clarity.

A process-driven infographic in beige and pink titled "Process and Business Value (ROI)." The top section highlights statistics from McKinsey and Forrester showing that design-driven companies see higher revenue and a $100 return for every $1 invested in UX. Below, a five-step timeline illustrates the UX GIRL workflow: starting with Clarity (research), moving to Structure (wireframes) and Vibe-Ready Design (modular UI), then to Accelerated Vibe Coding (AI-assisted dev), and finishing with a Live Product delivered in weeks instead of months.

How UX GIRL Designs for the Vibe

At UX GIRL, we design products that are dev-ready from day one. We don’t just deliver pretty interfaces - we deliver structured UX logic, scalable UI systems, and ready-to-deploy design blueprints that flow with your dev process.

Our process starts with UX workshops and research. We define user goals, create flows, build wireframes, and then bring it all to life in pixel-perfect UI. But here’s what makes us different: we design with vibe coding in mind.

That means:

Components are modular.
Layouts are logical.
Interactions are intuitive.
Everything is built to accelerate fast development and AI-assisted workflows.

Design for Vibe Coding is perfect for startup teams, AI-powered dev teams, no-code/low-code builders, and fast-scaling CTOs who need to ship fast - without sacrificing quality.

From Strong UX to Beautiful UI - in Record Time

Your MVP doesn’t start with code. It starts with clarity. A strong UX foundation and ready-to-use UI allows you to build smarter, faster, and better - whether you’re working with a team of devs or solo coding with AI.

With UX GIRL, our clients go from concept to implementation in weeks - not months. Our design packs are crafted to minimize development delays, boost usability, and drive adoption from day one.

Ready to Vibe Code?

If you’re building a product fast - and want it to work beautifully - start with a design that fuels your flow. With Design for Vibe Coding by UX GIRL, you’ll go from idea to live product faster than ever.

👉 Let’s design your next product the vibe way. Contact UX GIRL today.

‍

In a world where every sprint is a race and every pixel matters, the idea of artificial intelligence stepping in to support UX work is no longer theoretical. Zeplin -long trusted by product teams for streamlining design-to-dev handoffs - is entering new territory with its latest feature: AI Design Review.

But is it a true productivity boost or just another buzzword?

At UX GIRL, we took a deep dive into how this feature works, who benefits from it, and whether it's ready for real-world product workflows.

What Is Zeplin’s AI Design Review?

Zeplin’s AI Design Review uses a combination of large language models (LLMs) and visual analysis algorithms to automatically evaluate design screens uploaded to the platform. With one click, the tool can assess:

Adherence to UI/UX best practices (contrast, visual hierarchy, spacing),
Alignment with your design system,
Text readability,
Accessibility issues,
Element alignment and padding.

The result? Designers and teams receive instant feedback before development even starts - saving time, reducing rework, and improving consistency.

Note: The feature is currently in Beta and available for Team and Organization plans

How It Works in a Real Sprint

Let’s say a designer finalizes a set of screens. Normally, they’d hand it over to the team for manual review, often leading to rounds of feedback, corrections, and delays.

With AI Review, here’s how it looks:

The designer uploads the file to Zeplin.
They trigger AI Design Review, which instantly scans layout, color, type, spacing, and accessibility.
The AI suggests corrections like:
- “Low contrast between button and background.”
- “Heading typography breaks consistency with subheaders.”
The designer adjusts accordingly.
PMs and developers receive a cleaner, more polished file with fewer errors.

Zeplin claims the feature can reduce design-related issues passed to development by up to 30%

Why Product Teams Should Care

For Product Owners, Project Managers, and CTOs, AI Design Review can:

Accelerate iteration cycles by catching issues early,
Improve design consistency across large or fast-changing UI systems,
Help non-designers (like PMs) understand design quality without relying solely on design reviews.

Future updates will allow teams to customize review rules based on internal design systems, making the tool even more relevant for enterprise environments.

Limitations to Watch Out For

Despite its promise, Zeplin’s AI Review isn’t a silver bullet - and it shouldn’t be treated as one. Key caveats include:

Lack of design intent: The AI can't understand why a designer made a specific decision.
No user context: It doesn’t analyze user goals, flows, or emotions behind the interface.
No support for non-English reviews (as of now).
Risk of over-automation: Teams may over-rely on AI and deliver “technically correct but uninspired” UI.

This makes it a great supporting tool, but not a replacement for thoughtful human review.

How to Integrate It Into Your Workflow

To make the most of AI Design Review, we recommend the following integration model:

Designer finishes a screen and uploads it to Zeplin.
AI Review is triggered, and suggestions are considered.
Project Manager reviews AI feedback before sprint planning or handoff.
Developers get cleaner, AI-reviewed designs, reducing back-and-forth and rework.

This model works best when combined with traditional team review sessions and design QA.

Is It Worth It? Our Verdict

If your team:

Moves fast (Agile, CI/CD),
Uses a design system,
Delivers at scale or across multiple platforms,

...then AI Design Review can help reduce errors, align expectations, and deliver better experiences faster.

Smaller teams may find it a "nice to have," but even then - testing it in one sprint can offer real insights.

Conclusion: So, Should You Use It?

Zeplin’s AI Design Review isn’t about replacing designers - it’s about giving them better tools to work smarter. It acts as a second pair of eyes, offering clear, structured feedback before handoff.

At UX GIRL, we help teams like yours build processes that combine AI-powered tools with UX strategy, research, and design expertise. If you’re curious about bringing AI into your workflow, let’s talk - we’ll show you how to make it work without losing the human touch.

‍

Contact us

Privacy Policy

GDPR compliance

About UX GIRL

When we collect personal data about you

Why we collect and use personal data

Type of personal data collected

Information we collect automatically

The use of cookies and web beacons

How long we keep your data

Your rights to your personal data

Hotjar’s privacy policy

Sharethis’s privacy policy

Changes to this Privacy Policy