Most companies approach AI completely backwards. They start with technology capabilities and hunt for problems to solve, asking “How can we use AI?” instead of “What problems do we actually need to solve?”

As someone who built and scaled Klipboard from startup to acquisition by Kerridge Commercial Systems in 2024, Draven McConville is someone who has firsthand knowledge of how the wrong approach to AI can waste resources and disappoint customers.

Having implemented AI-driven solutions that now serve thousands of field service businesses globally, I’ve learned that the numbers tell a sobering story: 42% of businesses scrapped most of their AI initiatives in 2024, up from just 17% the year before, with MIT research showing 95% of generative AI pilots failing to achieve rapid revenue acceleration.

The successful AI companies aren’t the ones with the most sophisticated models. They’re the ones that started with customer problems and worked backwards to find solutions.

The Technology-First Trap

Walk into any boardroom discussing AI and you’ll hear the same conversation. Someone mentions ChatGPT’s capabilities, another talks about computer vision, and suddenly everyone’s brainstorming possibilities without considering real business needs. Yet studies consistently show that 70-85% of AI projects fail, with failure rates double that of traditional IT projects.

Despite businesses surging AI spending to $13.8 billion in 2024, the vast majority of these investments aren’t delivering results because they start with technology rather than problems.

When GitHub Copilot became available, we could have implemented it everywhere. Instead, we asked: “Where are our developers spending time on repetitive tasks that could be better spent on customer-facing features?” This problem-first thinking made the difference.

Real Problems vs Imaginary Opportunities

The difference becomes clear when you talk to customers instead of just internal stakeholders. Your customers aren’t asking for machine learning algorithms. They’re asking for faster responses, more accurate information, and reliable solutions.

Poor data quality, inadequate risk controls, escalating costs and unclear business value are the primary reasons projects fail – all symptoms of starting with technology rather than customer problems.

At Klipboard, our field service customers had a specific problem: technicians spending too much time on paperwork. We could have built an AI system for voice-to-report generation. But their real problem was simpler: they needed software that worked offline, had large buttons for gloved hands, and could complete common tasks efficiently.

Our solution came later with automatic service request categorisation, routing requests to the right technicians. This wasn’t about implementing AI for its own sake. It was about solving defined customer frustration.

The Questions That Actually Matter

Instead of “How can we use AI?” ask:

• What are customers complaining about most frequently?
• Where do employees spend time on repetitive, rule-based tasks?
• What decisions could benefit from better data analysis?
• Which processes break down when volume increases?

Spend time with customers and front-line employees understanding their daily frustrations. Start small with clearly defined problems and measurable outcomes. Winning AI programs earmark 50-70% of their timeline and budget for data readiness rather than model development.

As AI becomes more accessible, competitive advantage won’t come from having the best technology. It will come from understanding which problems are worth solving and how to solve them effectively.

Origianl Creator: Ekaterina Pisareva
Original Link: https://justainews.com/blog/why-your-ai-strategy-is-backwards/
Originally Posted: Fri, 05 Sep 2025 16:11:49 +0000