AI Alone Won’t Magically Unlock Earnings Power

Artificial intelligence has quickly become a hot topic around dinner tables and in corporate boardrooms. But delivering business benefits from AI will take time. Investors should proceed with caution.  

There’s been an explosion of hype around the disruptive capabilities of generative AI. Since ChatGPT launched, AI’s popularity on Google has surged (Display, left). Everyone is playing with chatbots, but AI isn’t just for fun and games. Ask ChatGPT how AI might transform business, and it spits out a list of applications from virtual assistants for customer support to predictive analytics, fraud detection, and futuristic autonomous vehicles and drones.

Companies Rush to the AI Party

During first-quarter earnings season, about 20% of US companies and global companies talked about AI on their earnings calls, based on our parsing of earnings-call transcripts (Display, right). Unsurprisingly, technology companies were the biggest advocates, yet in consumer discretionary, financials, healthcare and industrials, AI was on the agenda, too. The fervor hasn’t infected every sector. But this revolution is just beginning.

AI requires massive computing power and, so far, its enablers have been the biggest winners in equity markets. Other companies that will be users of the technology are exploring how to deploy AI to solve bottlenecks and create efficiencies.

Don’t Be Seduced by Big Talkers

After seeing shares of AI enablers surge, investors might be enchanted by visions of invisible robots that magically unlock profitability. Yet we believe that companies must show—not tell—how AI fits into a business model. They’ll need to prove that the technology works reliably, is embraced by customers, improves productivity and supports earnings. If an AI technology becomes commoditized, its competitive advantages could get eroded. And keep in mind how many of the early dot-com darlings that promised to change the world disappeared without a trace.

Equity investors shouldn’t jump blindly on the AI bandwagon. AI isn’t an end in and of itself; it’s all about the applications. The challenge is to figure out how AI fits into different industries and investment theses, by asking which companies will benefit and what types of jobs are at risk. Manual repetitive desk jobs that require little innovation are vulnerable, and we’re already seeing chatbots that perform well when drawing from fixed information sets. Forward-looking companies might use AI to improve productivity, but don’t assume that every declaration about AI will deliver real business advantages. Companies that experiment and fail fast may actually find the best applications more quickly.

Data science can help investors begin to sort the strategic thinkers from publicity seekers. By asking the right questions as part of a fundamental research process, equity investors can then identify the truly innovative companies that will successfully plug AI applications into a broader business strategy to ultimately enhance investment returns.