Prioritizing customers while chasing the bleeding edge of generative AI

As AI presents a world of unlimited possibilities for software companies, it also has the potential to open Pandora’s box, unleashing unintended consequences. This is especially true as the feature set of AI approaches artificial general intelligence (AGI).

However, it can also be used to create useless hype that does not move work or leisure forward significantly. Innovation is a wonderful thing — but the question to confront is “Are we creating what customers really want?”

Pew Research Center revealed that 52% of Americans feel more concerned than excited about the increased use AI. There is a general unease that comes from media coverage, and also a lack of understanding of how this will impact people’s day-to-day lives.

Meanwhile, companies are frantically creating and shipping a lot of ill-conceived AI features their customers don’t adopt, driven by the fear of missing out (FOMO). Doing so not only costs more money — AI via API is not cheap — but also has the potential to undermine a company’s reputation.

In addition, building tools that customers don’t want and won’t use takes valuable time and resources away from investing in other features that may make a difference — a missed opportunity for the company and customers alike. 

While we develop innovative solutions, we can’t lose sight of our customers’ core needs today. AI innovation can do wonders, but to satisfy all customers from early adopters to laggards, from casual users to experts, companies need to navigate the balance between hype-y innovation and simple, customer-grounded solutions.

Tech reality check: Diffusion is slower than you think

AI, especially the new, generative kind, isn’t hype — it’s real, it’s advancing, and it’s here to stay. That said, technology companies and vendors often overestimate the diffusion rate of technology into the real economy. It’s much slower than we think. 

For example, the first smartphone was released in 1993, but the first iPhone didn’t debut until 2007. In 2011, only 35% of Americans owned a smartphone. It took a decade for that figure to reach 85%, and it’s taking the rest of the world much longer. While it seems smartphones rapidly took over technology, communication and daily life, it has been over 30 years since the first smartphone was released and 16 years since the iPhone brought the full internet to our fingertips. 

There’s no denying that incredible technological breakthroughs have happened rapidly over the last few years, but that doesn’t mean people will adopt these innovations at the rate we wish or anticipate. Consider a recent survey by Morgan Stanley, which found surprisingly low usage of AI chatbots with only 19% of respondents having used ChatGPT.

So even when it feels to people in technology creator enclaves that “everyone is doing it,” many people on Main Street take their time to adopt the latest and greatest. 

Now that’s not to say that the tech isn’t incredible, but rather that the general public takes time to adopt new innovation at scale — think of it like a technology absorption rate. It took decades before mainframes took off in major businesses in the U.S., and decades for PCs to become ubiquitous.

While the cycles of technology adoption have since accelerated — it took much less time for mobile phones and the internet to proliferate — the time of adoption is still significant if the mobile revolution is anything to go by.

People, even knowledge workers, tend to adopt along four dimensions:

• Cost: If it’s too expensive, mainstream customers will stick to cheaper, tried and true technologies, unless they are in highly competitive sectors.
• Friction: New technology is often just something else to learn, and most people are not early adopters by definition. The bar has to be very low for new tech to become compelling.
• Availability: Non-early adopters will only venture when the technology is seen as ubiquitous and unavoidable.
• Reliability: The new technology has to be reliable. Most consumers are not interested in troubleshooting things and figuring out why they don’t work as intuitively as their existing toolset.

The fact that some of today’s greatest technological breakthroughs, like GPT-4 and the ecosystem it powers, are taking time to become mainstream, emphasizes that while these tools are in their early stages, many people will still resort to the tried and true ways they’ve always done things. Traditional search has hardly collapsed in the time that ChatGPT and Bing have integrated them into the search user interface.

Thoughtful AI innovation

What does this mean as you look to deliver new technology like AI-driven features? It means keeping your customers at the center of your innovation. 

Different people have different needs. Some will be early adopters, ready to try anything you ship to them. Some will be laggards looking for the simplest on-ramp and unwilling to be at the bleeding edge unless forced. There is additionally an expert spectrum — most customers are unsophisticated and crave simple intuitive interfaces for very powerful technology, but others are experts that crave a lot more creative control over the fine-tuning of the technology they purchase.

While it may seem some of these audiences are at odds, one thing all customers can agree on is that they want a tailored customer experience that is simple and intuitive, as well as powerful. 

AI should not be implemented simply to get credit for shipping something cool, that’s a pyrrhic effort. Instead, think of AI as a new tool to deliver customer-centric feature sets that can accelerate workflows or increase capabilities given the same workflows. For example, an AI copywriter can write four times faster than before or add illustrations that was otherwise impossible.

In addition, AI can unlock brand new super workflows. Instead of retrofitting AI into existing workflows alone, reimagine how AI can redefine the workflows entirely. Don’t just transcribe your Zoom meetings, but use recordings as a management tool to check on your team’s stress level through voice and sentiment analysis, or drive proactive task management by automatically adding follow-up tasks to your calendar.

The importance of ‘thoughtful AI’

Overall, it’s important to bring customers into the creative process to help co-create the application of these new capabilities. Tailoring the AI-enhanced software journey closely to pressing customer problems is critical to delivering thoughtful artificial intelligence-augmented feature sets.

“Thoughtful AI” is all about using AI as a core building block to deliver features that are simple, usable and powerful that your customers will return to again and again because they find them delightful and useful. Alternatively, AI features that are good announcements but offer little usable workflow acceleration and don’t get customers to rely on them over and over again generally don’t match how customers work today or how they want to work tomorrow.

Technology should adapt to humans, not the other way around. This simple philosophy should guide product innovation, and while it may seem like a tall task to provide products that work for all customers, the thoughtful application of AI can help strike the balance.  

We are in the midst of an innovator’s dilemma, an inflection point. Companies need to stay on the bleeding edge and satisfy the needs of early adopters with new AI-native workflows. However, in this quest to innovate, you can’t lose touch expressing the benefits of AI in a simple way that wraps around how your customers work every day.

As you reflect on how to prioritize your product innovation, my advice is this: Don’t just execute on your ‘technology ideas.’ Build new features around how your customers work that are enhanced by these new tools. Better yet, co-create them with your customers. Their pain points and your knowledge of AI’s possibilities will unlock new ideas that will be impactful.

Know what your customers need today, and anticipate how new, innovative approaches will result in the solutions for tomorrow. 

Oji Udezue is chief product officer at Typeform.