Layoffs at Watson Health Reveal IBM’s Problem With AI

Axed engineers say IBM isn't always smart about artificial intelligence

Illustration of IBM logo with stethoscope.
Illustration: IEEE Spectrum; Images: IBM; iStockphoto
Advertisement

IBM, a venerable tech company on a mission to stay relevant, has staked much of its future on IBM Watson. The company has touted Watson, its flagship artificial intelligence, as the premier product for turning our data-rich but disorganized world into a smart and tidy planet.

Just last month, IBM CEO Ginni Rometty told a convention audience that we’re at an inflection point in history. Putting AI into everything will enable businesses to improve on “an exponential curve,” she said—a phenomenon that might one day be referred to as “Watson’s Law.”   

But according to engineers swept up in a major round of layoffs within IBM’s Watson division last month, the company’s promotions of its “cognitive computing” platform mask its own real difficulties in turning its AI into a profitable business. 

“IBM Watson has great AI,” one engineer said, who asked to remain anonymous so he wouldn’t lose his severance package. “It’s like having great shoes but not knowing how to walk—they have to figure out how to use it.”

The layoffs at the end of May cut a swath through the Watson Health division. According to anonymous accounts submitted to the site Watching IBM, the cuts primarily affect workers from three acquired companies: Phytel, Explorys, and Truven.

All three companies, acquired in 2015 and 2016, brought with them hefty troves of health care data as well as proprietary analytics systems to mine the data for insights. The companies also brought existing customers: health care providers that used the analytics to improve both their care and their finances. 

Two laid-off engineers from Phytel spoke to IEEE Spectrum in depth. They allege that IBM’s leadership mismanaged their company since its acquisition, and say the problems at Phytel are emblematic of IBM’s struggles to make Watson profitable. Several other Phytel employees corroborated the basic facts of their accounts.

Both engineers worked for Phytel since before its 2015 acquisition, and say they were excited to become part of Big Blue. “Everyone expected that we would join IBM and be propelled by their support, that it would be the beginning of great things,” says the first engineer.

“Phytel was a really good company, it was really strong,” says the second, who also asked for anonymity to preserve his severance package. “Many people who got laid off are disappointed because they helped build this company and they really thought IBM would take it to the next level—not destroy it within three years.”

Phytel had more than 150 clients when it was acquired by IBM, the engineers say. Today it has about 80. “Smaller companies are eating us alive,” says the first engineer. “They’re better, faster, cheaper. They’re winning our contracts, taking our customers, doing better at AI.”

“A small percentage of the workforce”

IBM spokesman Ian Colley has declined to answer IEEE Spectrum’s questions about Phytel, and won’t comment specifically on the engineers’ allegations.

Instead Colley provided the following statement about the layoffs: “This activity affects a small percentage of our Watson Health workforce, as we move to more technology-intensive offerings, simplified processes, and automation to drive speed. IBM is continuing to reposition our team to focus on the high-value segments of the IT market, and we continue to hire aggressively in critical new areas that deliver value for our clients and IBM.”

Colley also references a Morgan Stanley analyst’s note, which argues that the layoffs represent restructuring that typically follows an acquisition.

But not all financial analysts have been bullish on IBM Watson. Last July, a report from Jefferies LLC argued that despite IBM’s heavy investment in Watson (which the analyst estimated at US $15 billion from 2010 to 2015 alone), the division won’t be profitable.

Even though IBM Watson “remains one of the most complete off-the-shelf [AI] platforms available on the marketplace,” the analyst wrote, customers face serious difficulties in integrating Watson with their existing data and analytics systems. As an example, he cited the MD Anderson Cancer Center in Texas, which suspended its Watson project after spending $62 million on it. 

The Jefferies analyst also argued that IBM is “outgunned in the war for AI talent and will likely see increasing competition”—a situation that probably won’t be helped as disgruntled engineers leave IBM Watson with tales of mismanagement. 

Health care: the killer app? 

IBM Watson first wowed the public in 2011 with its triumphant performance on the game show “Jeopardy!” The AI owed its victory to its natural-language processing abilities, which enabled it to parse the show’s complicated and pun-filled clues, then scan its vast library for the solution. 

Two years after that televised tour de force, IBM declared that Watson would find its first killer app in health care. It seemed a natural fit: With hospitals and doctors switching to electronic health records for their patients, data troves were becoming available to machines like Watson.

For its first job, Watson was supposed to go to work at cancer treatment centers, scanning the medical literature and patient records to advise doctors on treatment plans. Three years later, an investigation by the publication Stat alleged that Watson Oncology hadn’t lived up to expectations. 

Over the years, IEEE Spectrum has reported that IBM was investing billions in the Watson business but that “Doctor Watson” wasn’t yet revolutionizing medicine. We didn’t neglect its significant achievements, though, also reporting on a study in which Watson came up with a treatment plan for a brain cancer patient in 10 minutes—as compared to the human experts’ 160 hours of deliberation. 

“Everything stopped”

IBM acquired the three companies Phytel, Explorys, and Truven for their databases of patient information and their data analytics, which helped physicians and insurance companies provide more effective care. (A fourth acquisition, Merge, brought in analytics relating to medical imaging, a very promising field for AI.) 

Watson Health hoped to use these assets to help health care systems switch over from the traditional fee-for-service model to the new value-based model emphasized under Obamacare. In the new model, health care systems aren’t paid for each procedure and visit but are instead rewarded for keeping patients healthy and achieving better outcomes. 

Phytel’s contribution was analytics paired with an automated patient communication system. A clinic could use the system to search its patient records and find, for example, all the men over age 45 who were overdue for a colonoscopy, and then use an autocall to remind them to schedule the dreaded appointment. Finding these patients who are “out of compliance” is a key part of preventative care, in which regular checkups can catch problems early, before they’re difficult (and expensive) to treat. 

Phytel came to IBM as a successful business with a great reputation. In a 2014 survey of companies working on population health management, the research firm KLAS ranked Phytel as No. 1 and called it the “clear leader” in the market. 

In a 2015 press release announcing the acquisition, IBM announced that Phytel would be part of its brand new Watson Health division, which would augment Phytel’s offerings with IBM’s “powerful cognitive computing.” 

After the acquisition, IBM management started the process known internally as “bluewashing,” in which an acquired company’s branding and operations are brought into alignment with IBM’s way of doing things. During this bluewashing, “everything stopped,” the first Phytel engineer says, and the workers were told not to focus on improving their existing product for current clients. “People were sitting around doing nothing for almost a year,” the second engineer says. 

Instead, they were instructed to work on a grand project of merging Phytel’s databases and analytics with those of Explorys. The integration was a massive task, the engineers say, which required a huge investment of time and money from IBM.

Coding without a map

Both engineers lay the blame squarely on IBM’s “offering management” department, which is responsible for coming up with the products that IBM sells to customers. 

IBM’s goal, the Phytel employees said, was to create a fancy new product that combined the capabilities of Phytel and Explorys. However, the offering managers didn’t have a clear idea of what that product would be. “They couldn’t decide on a road map,” says the second engineer. “We pivoted so many times.”

Both Phytel engineers say the offering managers didn’t have technical backgrounds and sometimes came up with ideas for new products that were simply impossible. While the engineers tried to follow the shifting directions from above, Phytel didn’t deliver anything new to the market. “And we were burning through money,” the second engineer says. 

Eventually, the engineers say, the offering managers did come up with a plan for a new product that would utilize both the Phytel and Explorys resources. But they note that there was no talk of bringing real artificial intelligence into the product. Watson, in other words, was absent. “They’re not anywhere close to injecting AI into the provider space,” says the second engineer. 

IBM recently began demoing its new product for patient communication with potential international customers, the engineers say—but these potential customers weren’t impressed. Instead they asked for something resembling Phytel’s old system. 

Layoffs

IBM hasn't released the total number of Watson Health employees laid off in May, but the Phytel workers estimate that at least 300 people were given notice at their company, Explorys, and Truven. 

At Phytel, the engineers estimate that IBM laid off about 80 percent of employees—upwards of 120 people from engineering, sales, and project management. “These are key technical people, and people working hands-on with the clients,” the first engineer says. “It’s not that they’re trimming down administrative staff or getting rid of nonessential workers.”

The engineers say that Phytel’s office, in Dallas, will soon have only a skeleton crew to support the existing products. How effectively that crew will perform, however, is an open question. The engineers say they were trying to help several major customers meet government-defined deadlines, such as starting with a value-based care model in July. “Now, it will be a big surprise to me if this goes as planned,” the first engineer says. “These layoffs will heavily undermine these efforts. If you cut all these people out, who will build the products? Who will deliver in the end?”

The engineers see these cuts as IBM’s retreat from making products for health care providers such as hospitals, clinics, and doctors’ offices. They say that the employees who escaped the layoffs at the three companies work on products intended instead for insurance companies. 

The laid-off employees say they aren’t too worried about finding new jobs. Hiring season is coming in August and September, they say, when companies plan their budgets. 

While an AI revolution in health care may be inevitable, they don’t believe that IBM will lead the charge. While Watson Health was floundering, they say, competitors have been racing ahead. “IBM has great potential, but in modern business, you adapt or go away. Workers will go to other companies,” says the first engineer. “The winner will not be IBM.”

The Human OS Newsletter

Biomedical engineering in a biweekly newsletter. Expert insights into wearable sensors, big data analytics, and implanted devices for personalized medicine.

About the Human OS blog

IEEE Spectrum’s biomedical engineering blog, featuring the wearable sensors, big data analytics, and implanted devices that enable new ventures in personalized medicine.