The 5 things IBM needs to do to win in artificial intelligence

This time, the transformation IBM faces is much more difficult – for two reasons. The first is simply the scope and depth of the change. IBM is leading the charge for a radical (and therefore risky) transformation in its client companies: using automated expertise at scale to efficiently solve problems too large and complex for humans to conquer alone. The second is that IBM's cloud-based services will likely cannibalize three IBM pillars: hardware, software and data center services. In making such a bet, IBM will face much more resistance than Gerstner did, both internally (business units protecting their turf) and externally (investors condemning weak short-term results and future risks).

To start this transformation, IBM has to make a number of changes in thinking, practice and culture, which can be especially difficult for established companies. The kinds of change we are talking about come more naturally to the highly successful startups of the past two decades – Google, Amazon, Uber and Facebook among them – and to their investors. The progress of IBM Watson in healthcare is a great example and driver of the transformation, providing big lessons for IBM and for all large companies that need to change their own business – or watch competitors do it first.

Based on our research at the Watson Health unit, interviews with IBM executives and discussions with many large IBM clients, we believe there are five basic moves the company needs to follow in order to sustain success. We explain each one, pointing out what IBM has done so far and what it needs to do next.

Develop a portfolio of capabilities with different risk/return profiles

Smart venture capitalists invest in a portfolio of startups with different risk profiles and payback times. Some investments lead to dead ends; others need to be combined to increase the likelihood of success. Only a few achieve great success. To achieve higher returns, most successful venture capitalists analyze their portfolios to gauge this risk/return balance (or tension) and focus on the companies with the best prospects.

Track record so far: just over a year after announcing a one-billion-dollar investment to launch the Watson group, IBM established the Watson Health unit in September 2015. Watson Health began to grow through acquisitions and a new partner ecosystem. IBM acquired established companies (such as Merge, a medical imaging management platform, Explorys, with one of the largest databases, and Phytel, a provider of healthcare analytics services) that can deliver immediate ROI for IBM and its clients. To seize medium-term opportunities, IBM formed partnerships with hospitals (for example, Memorial Sloan Kettering, The Cleveland Clinic, MD Anderson), medical device companies (Apple, Medronic, J&J and Novo Nordisk) and chorionic villus sampling retailers to develop intelligent applications for the Watson platform. These applications are expected to become available online over the next five years. To seize long-term opportunities, IBM is investing in startups that are developing applications for the Watson platform, including Pathway Genomics and Welltok.

What to do next: IBM must continue to think and act like a venture capitalist, quickly assessing the effectiveness and ROI of each of its hundreds of partners and acquisitions against Watson Health's goals and the timeframe for achieving them. IBM's leadership must have the courage to quickly modify or even eliminate those that are not performing according to plan.

Acquire ideas, not just revenue

The most successful innovative companies buy startups and other private companies to help with their own transformation. The strategic rationale for specific acquisitions varies. In some instances, they buy a company because it has strong financial performance in addition to an innovative product. A good example is Google's acquisition of Nest. Other times, they make expensive acquisitions of money-losing startups with great intellectual property (IP) and management teams with a long-term vision. For example, Google and Facebook have made IP acquisitions in artificial intelligence and virtual reality.

Track record so far: IBM has developed the foundation of Watson's IP mainly in its research and development labs. The company has also been an active buyer of small companies (for example, Alchemy for text analytics) and large companies (for example, assets of the Weather Channel). As mentioned above, IBM acquired Explorys and Phytel and combined their IP into a new offering, Watson Care Manager.

What to do next: Even when buying private companies, IBM has traditionally preferred to buy those with significant current revenue rather than those that have only IP. This strategy is natural, since revenue reflects a mature business with established products and business models that can be improved. To succeed with Watson and Watson Health, IBM's acquisition strategy needs to keep finding the right balance between buying large companies with market dominance and greater revenue impact and many more small startups with differentiated IP and strong teams that have long-term impact. In addition, IBM's corporate marketing must refrain from getting caught up in the artificial intelligence hype and resist Wall Street's pressure for short-term results by explaining how innovations like Watson's take years to pay off.

Grow a startup culture and spread it to the rest of IBM

A company becomes large when it creates highly efficient and effective processes to generate supply in response to market demand. Maintaining these processes requires a certain kind of corporate culture, one that eliminates anything that distracts employees from executing business processes. Startups, on the other hand, are about experimentation, rapid iteration, brutally honest assessment and rapid adoption of winning practices. Traditionally, new businesses within a large organization are seen as startups and can be a major distraction to a well-oiled, efficient machine. IBM and other innovative companies must move beyond this old dichotomy, infusing the energy and mindset of startups throughout the organization.

Track record so far: IBM has a long history of incubating new businesses (for example, the AS 400 computer and the PC). The company set up Watson and Watson Health with this philosophy in mind. They have their own spaces, a separate identity and the autonomy to define their own culture. Both units are aggressively hiring employees with an entrepreneurial mindset. The Watson organizations have adopted open offices, creative titles and innovative recruiting practices – and the goal is to expand these ideas and practices throughout IBM. Company-wide training programs teach new hires about startup culture, focusing especially on the need for speed.

What to do next: As it grows, IBM Watson must protect the startup/innovation culture it is establishing with new employees and facilities. Inevitably, because of Watson's continued success, we expect many IBM employees from other divisions to want to transfer to the Watson business unit. While IBM keeps working to bring Watson's startup culture to the whole company, it must also ensure that internal transfers absorb, rather than threaten, the growth of the startup culture.

Hire outside leaders who can take the ventures to the next stage

Successful startups and their venture capitalists continually assess whether a management team can take the company to the next stage. They may often mix a startup team at different stages with outside executives. Although the tenure of CEOs of public companies has decreased dramatically, corporations continue to emphasize management team stability and internal promotion. Companies with new businesses often find leaders from outside. For example, Google recently hired a Hyundai executive to run its Google Car initiative as it moves from experimentation to commercialization.

Track record so far: IBM is making the right management moves with Watson. It used a team with startup experience to take Watson out of IBM's research and development lab and make it a separate business unit. More recently, it brought in former Philips Healthcare CEO Debora DiSanzo to run Watson Health and complement the management teams of Merge, Phytel and Explorys.

What to do next: As Watson's business model solidifies, IBM must continue bringing in outside executives with the right experience to leverage the ventures – even if these executives do not fit the traditional IBM executive DNA. We are very encouraged by the hiring of DiSanzo to lead Watson Health and hope that IBM will continue to judiciously hire from outside to augment its current executive team.

Focus on just a few pursuits, with conviction

Successful startups typically focus on exploring one idea at a time, but are ready to pivot if necessary. They know how to achieve product/market fit and which ideas to pursue. However, large companies often make big bets simply because they can, often for no other reason than that they have the resources to make them. Focusing on just one or two big efforts (moonshots) at a time may not seem to make sense to many large companies.

Track record so far: Watson has pursued a wide range of potential solutions for its platform, from call center applications to financial analysis and health diagnostics. Watson Health itself is pursuing many solutions in areas such as cancer treatment and personal wellness. For example, it is developing many versions of an Oncology Advisor system with a variety of medical centers, including MD Anderson Cancer Center and Memorial Sloan Kettering Cancer Center, to name just two.

What to do next: Watson's and Watson Health's experiments are necessary. But they are also expensive and carry high risk – the biggest risk being that they are draining Watson's resources. While IBM is listening to customers – one of the company's cultural hallmarks –, it also needs a dose of pioneering spirit, for example, taking a point of view about the future independent of customer input. Steve Jobs is the greatest example of this method, as he refused to use market research to design the iPhone. We would like to hear IBM say: “while we are pursuing dozens of applications for Watson and Watson Health with clients and across our ecosystem, we are pursuing these few opportunities based on our own insights and convictions”.

The Watson-led transformation of IBM represents another moment of truth for the ninety-two-billion-dollar, one-hundred-and-four-year-old company.

The path will be much more difficult this time than it was during Gerstner's tenure, and IBM needs to follow the techniques of startups and venture capitalists more closely. Only by learning from these companies – and truly embracing key aspects of the startup mindset and culture – will IBM be able to regain its place as the most dominant technology company in the world.