What We Can Learn from the AI Backlash

Every major technology wave follows a familiar cycle: excitement, over-investment, disappointment, and then steady value creation. Most recently we saw it with Cloud, and SaaS, and now AI.

The current backlash with AI isn’t because the technology doesn’t work. It’s because too many jumped headfirst into the tools without asking the most important questions:

• What problem are we solving?

• Who is the user?

• How will this fit into existing workflows?

• How will we measure impact?

Instead of grounding AI in real needs, many people assumed the technology itself would be the solution.

Lessons From the Cloud and SaaS Waves

When Cloud computing first emerged, companies raced to migrate workloads. But many discovered that lifting and shifting old architectures only ballooned costs without delivering agility. The real value came when businesses redesigned for cloud-native.

Netflix, for example, took seven years to fully re-architect for AWS, and the payoff was resilience and scalability. Airbnb leveraged elasticity to match traffic surges during peak booking seasons.

With SaaS, the same pattern repeated. Organizations subscribed to dozens of tools without integration or governance strategies. The result: tool fatigue, wasted spend, and frustrated employees. Companies that succeeded like Atlassian focused on how the technology could solve real collaboration and workflow pain points.

The lesson is clear: technology doesn’t solve problems by itself. User focused design, adoption, and context matter.

Why AI Feels Different (But Isn’t)

AI amplifies the cycle because it feels like magic. Leaders imagine it can do everything — automate decisions, generate strategies, replace entire workflows. But just like all other technology, AI needs careful design.

• No clear use case → pilots that go nowhere.

• No user research → tools employees won’t adopt.

• No evaluation of the real problem → solutions in search of a problem.

According to McKinsey’s 2023 AI Survey, only 23% of companies using GenAI report measurable bottom-line impact. The main blockers? Lack of integration, no clear strategy, and solutions built in isolation from users.

Companies Getting It Right

Despite the noise, some organizations are showing what good looks like:

• Cisco (Technology, Global): Cisco has applied AI to a pressing enterprise challenge — understanding application performance across hybrid and multi-cloud environments. Its AI-driven Full-Stack Observability (FSO) correlates data across apps, networks, and infrastructure, detecting anomalies and predicting failures. The result: reduced downtime, improved customer experience, and lower operational costs. This ties AI directly to measurable business value.

• Duolingo (Education, Global): Focused on a clear user need — making language learning more engaging. Their “Duolingo Max” integrates conversational practice and real-time explanations, directly tied to outcomes. Engagement and retention are higher for AI-powered users.

• NIRAMAI Health Analytix (Healthcare, India): Tackled a pressing issue — affordable, non-invasive breast cancer screening. Their AI system uses thermal imaging and ML to detect anomalies, with over 500,000 women screened across 150+ hospitals. By addressing a real problem in a culturally sensitive way, they achieved adoption where traditional solutions failed.

• Ping An (Financial Services, China): Embedded AI into health consultations and claims processing. Their “Good Doctor” app now has 300 million+ users, reducing hospital queues and broadening access to healthcare.

• Rakuten (Retail, Japan): Uses AI to personalize shopping recommendations and optimize logistics. During peak seasons, AI-driven demand forecasting has cut stockouts and improved delivery times, directly improving customer satisfaction.

• Grab (Super App, Southeast Asia): Integrates AI into ride-matching, fraud detection, and financial inclusion services. Their AI credit models enable microloans for underbanked populations, expanding access to financial services across the region.

• Rolls-Royce (Transport, Global): Embedded AI into digital twins for engines, enabling predictive maintenance. Airlines using this — including several across Asia Pacific — have seen 15–20% fewer unscheduled maintenance events, saving millions.

• TD Bank (Financial Services): Used AI to validate documents and reduce mortgage approval times from weeks to days.

• ServiceNow (Enterprise IT, Global): Reframed AI as “AI for workflow,” not “AI for everything.” Deloitte reported a 30% improvement in service automation with minimal retraining.

Each of these examples shows the same pattern: start with the problem, design for the user, and tie AI to measurable outcomes.

The Path Forward

These recurring success factors apply just as much to AI as is it did to Cloud and SaaS:

1. Defining Why. Shape the right value case to drive prioritization and momentum. AI for AI’s sake will not stick.

2. Planning to Succeed. Balance the need to move fast with a plan that accounts for complexity across business and technology.

3. Preparing for Change. Build a startup-like, customer-first culture that is ready to adopt new ways of working.

4. Transforming for Outcomes. Keep a holistic perspective — agility, scalability, and security must all be embedded, not bolted on.

And remember:

• Start with MVPs. Learn fast, fail fast, and continuously improve.

• Anchor in the customer. Maintain a user-first perspective at every step.

• Measure relentlessly. Tie outcomes to revenue growth, cost savings, risk mitigation, and customer delight.

This mirrors what MIT Sloan found in 2023: companies that applied human-centered design to AI were 3x more likely to see measurable ROI.

Final Thought

We must design better for AI. The backlash is a signal of the need to approach it with discipline, empathy, and design thinking. Those who pause to understand problems and user needs before chasing the next tool will be the ones who unlock AI’s lasting value.