Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

Obsolete on arrival: How to build new professional tech when the AI models keep changing

Product, design, and engineering teams face a new challenge: How can they effectively develop solutions based on AI models when the underlying technology may change within months?

Obsolete on arrival: How to build new professional tech when the AI models keep changing

[Courtesy of Charlie Claxton]

Profile photo of Fast Company Executive Board

Remember when ChatGPT was the hot new thing? Just over two years ago, all anyone could talk about were the game-changing effects that OpenAI’s product would have on technology, customer experiences, life…you name it. Since then, we’ve seen the launch of dozens of large language models—from open-source options like Llama 3.1 and Mistral, to premium general-purpose systems such as Claude 3.5 and GPT-4, to industry-specialized solutions like DeepSeek and BloombergGPT—and the list keeps growing.

Technology has always evolved, but now it’s happening at unprecedented speed. The window between introduction and widespread adoption seems to shrink every day, and this can make it overwhelming for companies to keep up. Product, design, and engineering teams face a new challenge: How can they effectively develop solutions based on AI models when the underlying technology may change within months?

NEW TOOLS, PERSISTENT PROBLEMS

Many companies make the mistake of chasing the latest AI tools without having a clear connection to customer needs in mind. Instead of solving existing pain points, they try to invent new problems to fit the available technology. This results in solutions that customers don’t actually want or need. Assuming that customers will drastically change how they interact with products just because a new AI model is available is a recipe for disaster.

Rather than getting caught up in the latest AI advancements, businesses should focus on solving the ongoing challenges their customers face. While AI models may change quickly, fundamental customer needs tend to stay relatively unchanged month to month and year to year. As Jeff Bezos has said: “You can build a business strategy around the things that are stable in time.” This is a good approach to take when designing with new technologies.

For example, legal professionals are always under pressure to save time without missing critical details and making costly mistakes. In the tax industry, firms often struggle with a lack of new talent and the burden of tedious, low-level tasks. These are real and ongoing problems, and the customers don’t really care how they’re solved—just that they are.

Once we understand those core needs, we can then look at the new models as tools in our arsenal. To effectively build AI-powered products, companies should start by answering two key questions. First, what does this technology actually do? And second, does it solve our customers’ real problems? Building a better experience might involve the latest AI or it might not. The goal is to address those pain points and create better experiences for our customers—and do it better than our competitors.

AI INTEGRATION AT THOMSON REUTERS

At Thomson Reuters, we’ve leveraged our 30+ years of AI expertise to prioritize customer experience over technology hype. This has allowed us to quickly map AI capabilities to real customer problems, providing a faster market entry than competitors who are still looking for use cases.

Our process begins by identifying customer pain points using the customer experience outcomes (CXO) framework, data analytics, and direct engagement. Once we clarify the friction points, we explore multiple ways to address them. AI is just one of many possible solutions, and if we do incorporate it, we make sure to offer varying levels of control so customers feel empowered instead of sidelined by automation.

For example, when we integrated AI into Westlaw, one of our flagship legal products, we designed the AI chat interface to work with the natural-language search patterns they were already familiar with. The AI improved the quality of results without requiring users to become expert prompt engineers overnight. By meeting customers where they are, we created a better experience without causing undue disruption.

Even as AI continues to change at a breakneck pace, one thing remains constant: In the end, customers care about getting their work done more efficiently and with less friction, and any details about how that happens are secondary. Companies that concentrate on solving customer pain points by delivering better experiences, not just the latest technology, will come out ahead.