By: Elena Mart

Saloni Pasad is a Senior UX Designer with an M.S. in Human-Centered Design and a Bachelor’s in Communication Design. She works with global clients whose platforms serve millions of users, where the stakes are high, and her creative design choices influence not just user experiences, but also the potential success of entire businesses.

Despite significant investments, many AI projects do not always succeed, and it’s rarely because the technology is fundamentally flawed. We spoke with Saloni to uncover what’s often missing in AI design.

Q. Why do so many AI projects fail, even when the technology is advanced?

A. Artificial intelligence is often viewed as a transformative tool capable of changing how we work, create products, and even shape our daily lives. However, in many cases, AI projects do not meet expectations,not because the technology is inherently bad, but because the creative thinking behind it often does not fully address the real-world challenges it seeks to solve.

From a designer’s perspective, the problem is seldom technical. It’s about human value: AI tends to be most effective when it focuses on solving problems that are meaningful to people. Yet, too often, projects prioritize attractive features or overhyped ideas that might seem innovative on paper but don’t truly resonate with end users.

Q. What’s missing when AI promises innovation but doesn’t deliver?

A. The core issue with AI’s creative struggle can be described as the “innovation gap.” Designers typically focus on desirability, what will delight users and improve their lives, while data scientists often emphasize feasibility, what the technology is technically capable of doing. The sweet spot, where AI is both practical and achievable, lies in the overlap of these two perspectives.

I’ve observed teams spend months developing clever AI features that are technically impressive, but ultimately fail to address meaningful problems. What feels innovative in the lab may not always translate well to real-world use. The result is often frustration, wasted time, and missed opportunities.

Q. Even when it’s not perfect, where can AI really make a difference?

A. AI does not need to be flawless to be effective. Some of its most useful applications enhance human creativity rather than replace it. For instance, Gmail’s Smart Compose can help you draft emails more efficiently, Spotify’s Discover Weekly suggests music you might enjoy, Google Photos curates vacation highlights, and Apple Photos lets you search for a ‘dog at beach’ to find that one photo.

These tools do not aim for perfection; rather, they assist, inspire, and speed up processes, leaving the final decisions to humans. Think of AI as a novice collaborator: eager but still learning, inconsistent at times, but surprisingly helpful in the right situations. You wouldn’t trust a beginner to fly a plane, but you might let them mix music at a party. Similarly, AI excels in low-risk, repetitive, or time-consuming tasks, freeing humans to focus on more complex, creative work.

Q. From your experience, what patterns cause AI projects to stumble?

A. From a design perspective, AI failures often follow predictable patterns:

• Some AI features fail because they do not address real user needs. 
• Tools that do not generate revenue, improve efficiency, or align with business goals are often deprioritized or abandoned. 
• When companies use scraped social media data or biased datasets to train their AI. 
• When AI models fail in real-world conditions and cannot deliver accurate predictions. 
• When AI feels unfair, lacks transparency, or does not center on human needs, it erodes user trust.

Q. How can we design AI that actually enhances creativity and human value?

A. The most compelling AI projects don’t chase technological breakthroughs for their own sake; they focus on human-centered problems that AI can meaningfully enhance. The criteria are straightforward:

• The AI solution should provide high user value. Does it make life easier, more enjoyable, or more creative? 
• It should be low-risk. Can people still succeed if the AI makes a mistake? 
• Its technical demands should be reasonable. Is the solution feasible without requiring perfection?

A tool that helps a designer organize assets or suggest creative ideas doesn’t need to be flawless. It just needs to perform better than the default, helping humans amplify their own abilities.

Q. Looking ahead, what’s the future of AI when we put humans at the center?

A. The next phase of AI innovation will likely come not from pushing models to their technical limits, but from rethinking how we design with human value at the forefront. It’s about identifying the right problems, embracing imperfection, and creating AI tools that work alongside humans, rather than trying to replace them.

AI won’t replace people; however, misaligned or poorly designed AI projects can limit creativity, waste resources, and even pose risks to the organizations that develop them. The key is fostering a creative partnership, not focusing solely on technical prowess. When AI and design meet at that ideal intersection, the results can be innovative, unexpected, and deeply human.