Why the AI Boom is Not Going According to Plan
The AI boom is not living up to expectations, as organizations struggle to turn their AI investments into reliable revenue streams. Despite initial excitement and high expectations, enterprises are finding generative AI more difficult to deploy than anticipated. Additionally, AI startups are being overvalued and consumers are losing interest. Even McKinsey, which initially forecasted $25.6 trillion in economic benefits from AI, now admits that companies need to undergo “organizational surgery” to unlock the full value of this technology.
The Importance of Starting with Product-Market Fit
Before rushing to rebuild their organizations, leaders should first go back to basics. Just like any other endeavor, creating value with AI starts with product-market fit. This means understanding the demand that needs to be met and ensuring that the right tools are being used for the task at hand.
Using Hammers to Cook Pancakes
In the current AI landscape, everything is being hammered. At CES 2024, attendees were amazed by AI toothbrushes, AI dog collars, AI shoes, and even an AI button on a computer mouse. In the business world, 97% of executives expect generative AI to add value to their businesses, with three-quarters of them handing off customer interactions to chatbots. However, this rush to apply AI to every possible problem has resulted in many products that are only marginally useful and some that are even destructive. For example, a government chatbot incorrectly advised New York business owners to fire workers who complained about harassment. Turbotax and HR Block also faced issues with their bots giving bad advice half of the time.
The Furby Fallacy
AI is uniquely prone to disrupting businesses’ existing processes for establishing product-market fit. When using tools like ChatGPT, it’s easy to be fooled by how human-like they seem and assume they have a deep understanding of our needs. This is similar to the Furby fallacy, where people believed the toys were learning from their users when they were actually just executing pre-programmed behavioral changes. The tendency to anthropomorphize AI models can lead to overestimating their sophistication. The Alignment Problem, which refers to the challenge of issuing precise instructions to AI models, makes establishing product-market fit even more crucial for AI applications.
Getting Back to Basics
Since AI systems cannot find their own path to product-market fit, it is up to leaders and technologists to meet the needs of customers. This involves following four key steps: understanding the problem, defining product success, choosing the right technology, and testing the solution. Many companies go wrong by assuming their main problem is a lack of AI and adding AI without considering the actual needs of the end-user. By clearly articulating the problem first, it becomes easier to determine if AI is the right solution and which types of AI are appropriate for the use-case. Defining what will make the solution effective is also crucial, as there are always trade-offs to consider. Once the goals are clear, working with engineers, designers, and partners to choose the right technology and address potential constraints early in the process is essential. Finally, testing and iterating the solution ensures that it meets real needs and creates real value.
Drawing Bullseyes and Hitting Targets
The temptation to deploy any AI application in any setting often leads organizations to “innovate” without a clear plan. This approach results in numerous arrows being fired off randomly, with bullseyes being drawn around the spots where they land. While some arrows may hit useful places, the majority will yield little value for businesses or end-users. To unlock the true potential of AI, it is necessary to draw the bullseyes first and then focus all efforts on hitting them. This may involve developing solutions that do not involve AI or using simpler and more focused AI deployments. Regardless of the type of AI product being built, establishing product-market fit and meeting customers’ wants and needs is the only way to drive value. Companies that prioritize this will emerge as winners in the AI era.
In conclusion, the AI boom is facing challenges in delivering on its promises. It is crucial for organizations to focus on product-market fit and avoid the temptation to deploy AI without clearly understanding the problem and the needs of end-users. The Furby fallacy highlights the danger of overestimating AI models’ capabilities and reinforces the need for precise instructions. By following a back-to-basics approach and prioritizing product-market fit, companies can unlock the true potential of AI and create real value in the AI era.