Few companies have put their weight into artificial intelligence (AI) as heavily as International Business Machines (IBM -0.01%). The company started its transformation into a cloud and AI specialist way back in 2012, when Ginni Rometty took over Big Blue's CEO chair. It has been a long and painful road since then, and the IBM you see today bears little resemblance to the celebrated one-stop tech shop it was before Rometty's strategy shift.
So IBM started building its AI expertise years ago, infusing its machine learning tools into most of its products and services. And then OpenAI came along, raising the public profile of AI tech with its innovative ChatGPT system.
Will IBM's long-standing commitment to AI start to pay off in this updated market, or was Big Blue building computerized castles in the cloud without a clear path to profitable growth?
A consistent AI philosophy
First, IBM is under new management, but that doesn't take away from its AI focus at all.
Arvind Krishna, IBM's CEO since April 2020, managed the company's cloud and cognitive software operations for five years. Krishna is deeply committed to using, developing, and selling AI, and his years of high-level responsibility in this field should give him robust insight into how AI actually works.
As the driving force behind IBM's $34 billion buyout of open-source software veteran Red Hat, Krishna clearly isn't afraid of taking the business strategy in radically different directions. The current IBM business has been described as a larger-scale version of Red Hat, aiming at a $1 trillion hybrid cloud opportunity.
His strategic decisions underscore a commitment to AI and cloud technologies as core elements of IBM's future growth.
Lessons learned along the way
At a recent tech industry conference, Chief Commercial Officer Rob Thomas outlined how far IBM's Watson systems have come in the last three years:
We've been investing [in the WatsonX generative AI platform] for 3 years. And if you think about the arc of Watson, the first instance was really around machine learning, deep learning. We had some successes, we had some failures. It was really hard. You think about those projects, a lot of it was data labeling, data annotation. It was so much manual effort that you had some success, you had some failures.
In Thomas' view, the availability of powerful generative AI systems such as OpenAI's GPT and IBM's WatsonX can be a game changer for sophisticated deep learning systems. With the help of automated data labeling, WatsonX can train its supervised machine learning systems on previously unlabeled data -- a resource that's often in abundant supply for any company that collects real-world data through physical sensors or user input.
Image source: Getty Images.
Accuracy above all else
You won't see Big Blue launching consumer-friendly AI tools in a direct challenge to ChatGPT. Watson was always aimed at business-to-business services. The system comes with a plethora of fine-tuned learning models for specific business cases, and clients can always build their own Watson-powered AI tools around their own proprietary data sets.
And at the end of the day, there is no room for inaccurate AI ramblings when you're trying to run your business with these tools. So bulletproof accuracy is a top priority here, which isn't always the case when you're looking at the entertainment value of more creative AI systems.
Thomas continued:
We're really focusing on more narrow domains, things that we have the data to train models on. So things like cybersecurity, IT automation, digital labor, customer care. These are narrower use cases where we can build and train a model using IBM data. And then we go to a client, and we say, you bring your proprietary data, then that becomes your model. We think these will be very competitive. And also very accurate. And I think the most important distinction is, sometimes you get hallucination in consumer AI, there's really no appetite for hallucination in the enterprise. So accuracy is really important.
IBM just might stand on the threshold of a golden AI age
There is no substitute for actual business results, so investors will see how IBM's long-running AI focus plays out over the next few years. The generative AI cat is out of the bag, and you will see lots of different approaches producing wildly different business results as the AI market heats up.
If you bought into IBM's long-term AI and cloud computing many years ago, this is where the rubber should finally hit the road. The ChatGPT boom has opened everyone's eyes to what generative AI and deep learning tools can do for a data-heavy company. Big Blue needs to make the most of this unique opportunity, armed with a uniquely experienced and battle-tested portfolio of AI services.
Again, this potential overnight success was years in the making and IBM sacrificed a lot to get here. You might recall that pretty much every IT giant worth its salt wanted to copy IBM's success story in the early 2010s, and that's exactly where the company made a sharp turn into the current AI and cloud focus.
IBM Total Return Level data by YCharts
Investors are about to find out whether the long-term strategy shift was worth the pain. The stock has lagged the broader market since Rometty first took the helm, setting it up for either a final fade or a rambunctious rebound as the AI market evolves.
I expect IBM's deep AI expertise to start paying dividends at this point. The stock trades at the humble valuation of 13 times forward earnings. Speaking of dividends, it also comes with a generous 5.1% dividend yield. You shouldn't bet the proverbial farm on Big Blue today, but the stock could be worth a modest investment that motivates you to keep an eye on developments in the AI sector.
