What are the AI-proof skills every frontend developer needs?

Chances are you’ve heard the cliche by now: “AI won’t replace humans, but humans using AI will.”

It’s oft-repeated, but mostly because it rings so true.

AI and frontend development are tied at the hip. According to the State of WebDev AI 2025 survey, 91 percent of developers are using AI to generate code.

These AI tools are leveling the playing field. Almost any person, regardless of their coding knowledge, can now “vibe code” their way into a working application.

Naturally, the emergence of AI-assisted coding raises some questions.

If everyone is using AI to write code, how can the best developers stand out? How can you use AI better than everyone else? What are the skills you need to make sure you remain indispensable in the age of AI?

We gathered thoughts from seven senior developers to help provide some guidance:

• Jack Herrington, Software Engineer, Writer, and YouTuber
• Jemima Abu, Senior Product Engineer at CAIS
• Andrew Evans, Principal Engineer & Tech Lead, CarMax
• Lewis Cianci, Senior Software Developer, Queensland Health
• Emmanuel John, Lead Frontend Engineer, Gigmile Technologies
• Rosario De Chiara, Innovation Manager, Logistics and Financial Services Industry
• Marie Starck, Lead Developer, Fairstone Bank

Let’s see what each of them had to say about the skills you need to make yourself “AI-proof”:

Communication and empathy

Jack Herrington has worked in senior roles at some big-name companies, including Nike, Walmart, and most recently at Element Biosciences.

Communication with colleagues was always key, but Herrington feels that it’s even more important in the age of AI. At this point, AI code tools can’t replace a human developer’s ability to take technical issues and rephrase them in a way that non-technical colleagues can understand.

Herrington calls it “getting out of the world of the bits and the bites and into the world of what customers need – and the rapid iteration cycles around that.”

Lewis Cianci, a senior dev at Queensland Health, has written about the need for better communication and mentorship in the dev world.

“I’ve always thought that good developers are determined by their ability to communicate,” Cianci said.

This also applies to maintaining an AI-proof skillset.

“People don’t want to be jamming their stuff into AI and shipping the result; they still want someone to take responsibility for what’s being deployed,” Cianci added.

“Being a good communicator and being easy to work with are more important than ever.”:

Key takeaway: Learn how to communicate the meaning behind your code, for both technical and non-technical audiences.

An in-depth understanding of UX, accessibility, and system design

Accessibility is important in any industry, especially in financial services, where Jemima Abu operates with CAIS. She notes that while AI is great at coding grunt work, it hasn’t yet figured out the human side of UX and design. That’s where developers remain valuable:

“ When it comes to predicting errors or predicting user behavior or accessibility needs, that’s one area I don’t rely on AI for,” Abu said:

“I’m more likely to get in the code myself and be like, okay, what do I actually need to fix here?”

 

Emmanuel John’s team at Gigmile Technologies often uses ChatGPT’s UI generator and GitHub Copilot – saving hours of manual scaffolding of UI layouts or components.

What’s missing – and what devs can provide – is deep design thinking and testing.

“In-depth understanding of design systems and user experience is key,” John said.

“AI can generate components, but it lacks design taste and context. Being excellent at debugging, reasoning through weird behavior, and fixing what AI broke are huge skills.”

Key takeaway: AI hasn’t quite nailed the human side of UX, design, and accessibility. Make these core parts of your skillset.

Discipline in establishing boundaries for AI usage

Understanding when to use AI is just as important as understanding when not to, according to Andrew Evans.

“The ability to properly prompt tools is essential in working with AI,” Evans said.

“So is the ability to know when to do something yourself versus letting the AI do the work.”

Evans’ team at CarMax uses GitHub Copilot to conduct pull request (PR) reviews, and Claude as an “advisor” when learning new frameworks.

These tools help identify issues his team may have missed on first review, but they also make incorrect assumptions if they’re not fed the right context:

“In my day-to-day tasks, I often use an AI tool as an advisor, but I still will do the majority of the actual code creation. It’s important to be able to understand what you are building and not let the AI do too much.

Every developer should understand how to interact with AI tools, but also how to control what the tools do to your project deliverables.”

Key takeaway: Set boundaries on your AI usage so you can always maintain control over what you ship.

Interpreting AI outputs

At his logistics and financial services company, Rosario De Chiara and his team are experimenting with various AI IDEs as part of their daily development routine.

According to De Chiara, the best devs are able to parse through output from their IDE, identify what works, and – crucially – catch what doesn’t.

“We really need to invest time to understand code,” De Chiara said.

“Understanding what is generated is fundamental. We’re still in an age where what is generated can be fundamentally unmaintainable without previous planning.”

Key takeaway: Review AI outputs with real scrutiny. Don’t blindly accept its results without considering the greater goals of your project.

Problem-solving across functions

As a lead developer at Fairstone Bank, Marie Starck naturally collaborates with a lot of stakeholders.

As she sees it, the most AI-proof developers are the ones who can step outside the code and communicate across functions.

“Developers will have to rely more on their soft skills,” she said.

“My vision is that we will no longer need coders who purely code, but rather problem solvers who can sit at a table with product, engineering managers, and more to solve business issues.

We will always need technical people, but the developers who will succeed are the ones who have the right blend of soft and hard skills.”

Key takeaway Well-rounded devs are in demand in the age of AI. Focus on honing both your technical and non-technical skills.

---

AI isn’t going anywhere. Devs need to learn how to thrive alongside AI – whether that’s showcasing their problem-solving skills or finding the most creative, sustainable ways to use it.

Stay tuned for more insights from senior devs on AI and the future of frontend!