top of page
  • LinkedIn
  • LinkedIn
Pink Bubbles_edited.jpg

Why AI keeps missing the mark - and what it will take to deliver real value

April 2025 | Article

By Marianne van Ooij

Unlocking AI's full potential requires rethinking development workflows—aligning technology, UX, and organizational vision from the start.

The past year brought a surge of generative AI announcements from technology companies. More recently, however, we've seen a different kind of news cycle: stories of missed timelines, partial launches, and underwhelming rollouts. Publications including The Wall Street Journal, Hard Fork, and Bloomberg have highlighted how companies like Amazon and Apple have delayed, scaled back, or walked back their most ambitious AI feature promises.
 
This raises the question: if the technology is moving so rapidly, why does delivery seem to lag behind?

DOWNLOAD

Article

What AI is already capable of

To understand this gap, it's worth noting where AI has already delivered tangible value:

​

  • Large language models powering customer service, legal reviews, and code generation. Chatbots now resolve many customer queries; tools like Harvey assist legal teams; GitHub Copilot supports developers with real-time code suggestions.
     

  • Vision models detecting rare diseases, sorting inventory, and enabling creative tools. AI identifies eye diseases from scans; automates warehouse checks; and powers generative design capabilities in tools like DALL·E and Figma.
     

  • Predictive systems driving personalization, recommendations, and automation. Platforms like Netflix and Amazon personalize experiences, optimize pricing, and anticipate demand in supply chains.
     

These successes demonstrate that the technical foundation is in place. So what's holding back large-scale delivery?

The implementation gap: why AI promises fall short

Despite the power of the technology, many organizations still approach AI as either a feature to plug into an existing product or a purely technical problem to be solved by engineers and data scientists. This mindset limits how AI is integrated and what it can ultimately become.

​

Research (BCG, Rand Corporation) shows that technical issues account for less than 20% of AI implementation failures. The primary causes are strategic: unclear business objectives (38%), lack of user validation (25%), and organizational resistance (17%). Together, these factors contribute to an estimated $24B in wasted AI spending annually across Fortune 1000 companies.

Three critical gaps

Most AI implementations fail because of three critical gaps: the Vision Gap, the Experience Gap, and the Collaboration Gap:

 

1. Vision gap

Companies showcase ambitious AI capabilities in demos and marketing without aligning product development, engineering, and UX teams on implementation realities. Amazon's Alexa+ demonstrations highlighted complex capabilities like handling Uber orders and searching camera footage that weren't part of the initial rollout—creating a gap between expectations and delivery.

​

2. Experience gap  

AI systems introduce uncertainty, variability, and probabilistic outcomes that traditional product development processes aren't designed to handle. Apple's "More Personal Siri" features demonstrate this challenge—determining when an AI is confident enough to act requires design decisions about trust and risk, not just technical capability.

​

3. Collaboration gap

AI implementation requires seamless collaboration between previously siloed teams. Voice assistants illustrate this need clearly—they demand integrated work across natural language processing, system capabilities, and user interaction design to create experiences that feel natural and reliable.

​

These disconnects explain why even technically excellent AI systems often fail to deliver on their promise. The voice assistant examples highlight how AI requires more than technical excellence—it demands new approaches to product development that align vision, technology, and user experience from the start.

Getting real value out of AI requires transformation, not just new technology. It’s a question of successful change management and mobilization, which is why C-suite leadership is essential.  

​

— Alexander Sukharevsky Global co-leader, QuantumBlack by McKinsey

What's needed: a new operating model

AI represents not just a new technology but a new paradigm. Delivering its value requires reshaping the way organizations operate around AI initiatives:

​

  • A leadership vision grounded in clear problem definition, aligning innovation with practical execution

  • Understanding of both traditional and AI product development cycles

  • Cross-functional structures that support iteration, learning, and adaptability

  • New success metrics focused on outcomes, trust, and long-term engagement

  • Early integration of UX, ML, and product thinking

​

We've seen forward-looking organizations begin to embrace this shift—adopting new workflows, collaboration models, and ways of working where AI, UX, product, and engineering evolve together from the start.

​

This is precisely the kind of transformation our Four Shifts framework is designed to guide.

Successful AI integration: Spotify's Discover Weekly

 

While many AI features struggle to deliver, Spotify's Discover Weekly demonstrates what's possible with the right approach:

​

  • Cross-functional integration: Spotify formed teams that combined data scientists, engineers, UX researchers, and music experts from the start

  • Incremental value delivery: Rather than promising a revolutionary AI experience, they focused on solving a specific problem—music discovery—and refined it over time

  • Human-in-the-loop design: The system combines algorithmic recommendations with human curation and explicit user feedback mechanisms

  • Clear value proposition: Users understand what they're getting and why it's valuable, with appropriate expectations

  • Continuous improvement: The feature has evolved based on usage patterns and feedback, becoming more personalized while maintaining reliability

 

The result wasn't just a successful feature but a new model for how music discovery could be designed—driven by cross-functional collaboration and long-term strategic thinking—and a competitive advantage that has lasted years, not just made headlines

UX is not cosmetic—it's translational

AI systems are probabilistic, adaptive, and often opaque. Users need clarity, confidence, and control. These needs aren't new, but with AI's unpredictability and complexity, they've become more urgent. UX must now guide not just interfaces, but behavior, decision-making, and adaptation.

​

UX contributes to AI success by:

  • Creating pathways for users to understand, question, and influence AI decisions through transparent explanations and feedback mechanisms

  • Designing adaptive experiences that evolve with user needs

  • Spotting where models don't align with expectations or real-world context

  • Developing comprehensive metrics that track engagement, trust, and adaptation over time

​

We've described how this translational role becomes even more critical as AI capabilities advance. UX is how AI becomes usable, adoptable, and valuable—not just impressive in a demo.

Toward new workflows and partnership models

Success comes not from rushing to ship AI features, but from aligning leadership, design, engineering, and data science around shared goals. This alignment involves four key shifts that we've documented in our work at AI UX Navigator:

​

  1. Strategic Shift: Moving from deterministic mindsets to an AI-first strategy that embraces uncertainty and adaptation

  2. Research Shift: Evolving from episodic research to longitudinal discovery that tracks trust and adaptation over time

  3. Design Shift: Transitioning from static interfaces to flexible, probabilistic experiences with appropriate transparency and control

  4. Collaboration Shift: Transforming from silos to cross-functional teams with shared vocabulary and objectives

 

These aren't theoretical concepts—they're practical changes to how AI products get imagined, built, and improved. We've seen teams implement these shifts through tools like our AI UX maturity model.

Moving from hype to value

AI's early promise generated significant excitement. Its next chapter will be shaped by responsibility, integration, and intentional design.

 

Real value comes from organizations that assess their organizational readiness, develop a cohesive AI vision and strategy, and are prepared to change workflows and collaboration models to support AI implementation.

 

AI UX Navigator is committed to helping organizations navigate this transition thoughtfully. The path forward requires coordination between UX, technical teams, and leadership to create AI experiences that truly deliver on their potential.

To understand where your organization stands, we invite you to try our Organizational readiness assessment or explore our article about Successful AI implementation for practical next steps.

ABOUT THE AUTHOR(S)

Marianne van Ooij is the founder of AI UX Navigator

bottom of page