Eighty percent of AI projects fail, according to a recent RAND Corporation report. That is two times more likely to fail than other tech projects. However, after closing its next round of funding, ChatGPT—the startup that ignited the Generative AI frenzy two years ago—is predicted to receive a $100 billion valuation. Surprised? Be not.
Up to a point, the excitement surrounding new technologies typically endures even if they fall short of initial expectations. The Gartner hype cycle predicts that high hopes will be followed by a low point in disappointment. According to a June Gartner report, generative AI (GenAI) is most likely currently at this tipping point. This highlights the challenge of converting technology into engines of economic growth, rather than implying that the advancements in Large Language Models (LLMs) have not occurred. We just assume too much too soon.
According to Carlota Perez, a technology historian, the development of applications and modifications to organizational structures are always necessary for the next wave of technological innovation involving primary technologies, such as LLMs. For example, the development of electric motors and the reorganization of factory production lines to take advantage of these innovations were necessary before electricity gained significant traction. Companies can modify their AI adoption strategy by keeping this in mind.
First suggestion: Employ Google’s GenAI
How would you go about determining the distinction between GenAI and machine learning? Look it up on Google. You can then delve into the details by clicking on the list of links that Google gives you.
More recently, an AI-generated response in brief is also provided. This will be adequate most of the time. Initially, you can also ask the question in a GenAI application. One benefit of this is that it can spark a discussion in which you can raise more queries. Though it’s not usually your main concern, hallucinations can be a problem. In that case, you can always go into more detail later. In any case, learning is not a linear process.
Although this method of using GenAI is efficient, gradually becoming familiar with AI tools is a more significant but less evident benefit. You eventually discover which prompts work best and how to more accurately distinguish fact from fiction.
Additionally, you will discover the ideal uses for various tools. Some use it to take the place of relatively simple tasks that they used to outsource, like assisting them in drafting a press release or a simple legal document (one that is not very important). Others use it to generate fresh concepts, such as by looking for instances of related problems in other industries.
From an organizational standpoint, more ambitious AI integration into operations requires the widespread application of GenAI. People will resist using technology if they are uncomfortable using it. Full stop.
Second suggestion: See AI as a change initiative rather than a technology project
It’s simple to view AI mainly from a technical perspective. That’s a serious error. AI adoption necessitates new business procedures and behaviors. It is difficult to overcome the powerful force of inertia. It is only the first step to get people to feel at ease with a new technology in theory. A strategic transformation plan is required.
In The AI Playbook: Mastering the Rare Art of Machine Learning Deployment, Eric Siegel offers one. His first observation, using UPS as an example, is that people are typically more intimidated by large promises than inspired by them. Jack Lewis purposefully chose to first focus on assigning packages to trucks via deliver prediction when he began developing a system that forecasts tomorrow’s deliveries and suggests more effective delivery routes for drivers. Although it wasn’t as elaborate, it also needed less modification, which attracted senior management’s attention and made it simpler to execute.
Siegel’s second key finding is that labeling is important. If your AI project sounds ordinary, it will seem less menacing. That’s a good thing, since change can be scary. Once you’ve won, save the big words for your presentations. Start off by referring to “operational improvements.”
Ultimately, leaders who undergo change are best able to reap its benefits. Jack gave a skeptical UPS executive a ride on an algorithm-prescribed route, to which his first question had been, “So, are you working on anything important?” In an attempt to make a more efficient route, a counterintuitive turn was made at one point, leaving some packages that were nearby for later. At this point, the executive realized how much the new system could save costs.
Despite using this clever transformation strategy, it took years for UPS’s delivery system to fully incorporate AI. It now forms the core of the business’s entire optimization system. It saves $350 million in expenses and 185 million driving miles annually.
AI gains are not ending, but the hype is
Businesses will continue to reap the benefits of AI in general and GenAI in particular. However, as with all new technology, this is not a magic bullet, and the most difficult part is changing how things are done. The benefits of integrating AI into business operations are just getting started, even though the hype may have peaked.
