Not-So-Elementary Considerations For IBM Watson Services Buyers


Whether you have successfully started working with Watson, are evaluating it, did a PoC 18 months ago and swore off it, or have an enterprise license sitting around, you have realized that Watson is not your average prepackaged software application. As IBM’s umbrella brand term for all things cognitive, Watson capabilities range from analytics to cognitive solutions and virtual agents, available as individual APIs or prepackaged products to develop Watson applications. As cognitive technology like Watson falls in the “innovation” realm for most enterprises, it requires thorough experimentation, risk/opportunity assessment, project prioritization, steep learning curves on skills development, and above all, education and change management for the employee/customer base that is involved in the process.

When you’re working with a service provider through this journey, chances are they are on the same learning curve because of the newness of the cognitive market for business use. While IBM is taking Watson to market through its GBS organization, Watson APIs and products are being used by business and technology services providers in a variety of ways (see our POV paper on this subject here). IBM Watson technology has been around officially for a few years, and PoC projects are the norm so far. However, HfS hears a lot of industry optimism and “gearing up” for 2017-2018 being the years of more substantive implementations through this growing network of services partners.

Considerations for using Watson services

As you explore Watson in your organization:

  • Understand where and how your service provider is investing in Watson to offset cost: Perhaps the biggest barrier to Watson adoption for enterprises has been its high price tag for entry. Service providers have been trying to circumvent this by exploring options to host the Bluemix and Watson licenses plus external databases. Their clients can then access both the technology and data, particularly for proprietary solutions where cognitive APIs are being leveraged. Enterprises that already have access to the Bluemix cloud computing environment are getting started with Watson on it as an incremental investment. As Microsoft Azure, Amazon AWS and other competing cloud environments all have their own machine-learning technologies, the decision to which cognitive ecosystem you go with will likely be influenced by these larger technology-buying decisions.
  • Find the provider that Is collaborating with IBM in areas that matter to you: Watson APIs and products are being constantly revamped, retired, and regrouped and it will help to have advance knowledge from a service provider that is deeply involved with IBM in advancing specific areas. We heard instances of how by providing feedback to the IBM Watson product development team and working collaboratively, some service providers influenced the release of new functionalities that benefitted their clients’ projects directly. Look for the connections that your service provider team has been able to establish that could impact your particular use cases.
  • Find the Service Provider That Is Investing in Your Vision – Or Using Design Thinking to Help You Develop One: Even in these early days, we see industry, functional, and technological strengths developing among service providers. The experience gained and customization achieved with specific solutions – like Hexaware’s superannuation bot or Accenture’s mortgage advisor Collette – are valuable to companies that have already outlined these areas for Watson or are looking for new levers for value to their business and customer base. In areas where there is not a relevant standard solution, your leadership team will often have competing priorities. Consider service providers that offer Design Thinking workshops to establish the top business priorities, the process and technology roadmaps, and the definition of your own version of a future-state with an “augmented workforce”.
  • Don’t Underestimate the Power and Influence of Naysayers – Educate Them First: As shared by a financial services VP, “Internal stakeholders require fundamental lessons on what Watson is and isn’t…Our skeptics didn’t fully understand what cognitive or data mining benefits Watson brings; we should have expected it earlier on and addressed it head first”. Without aligning organizational buy-in, companies in our research have seen significant slowdowns in each stage of their projects. Make sure your key representatives understand the breadth of the technology and its suitability to your use case before kicking off and then check in regularly.

The ripple effect of Watson services 

With these considerations in mind, do note that whatever cognitive initiatives you undertake will invariably impact more than one part of the business and way of working. As a department undertakes the required data curation, reference architecture, process remodelling, and rollout, it will interact with and influence other departments or processes and advance their maturity toward more intelligent operations as well. For example, a retailer could go from a production pilot in personalized shopping on its website into cognitively determined best next actions for its sales channels, then on to cognitively driven merchandising and supply network on the back end to better predict demand. The core customer and product data can be leveraged across these functions and can become a powerful way to reinvent the entire customer engagement process.

The focus on better enabling the customer and/or stakeholder experience is driving significant enterprise interest to explore Watson services.

In our latest report, HfS Emerging Market Guide: IBM Watson Services, we further explore this theme of getting started with IBM Watson – the use cases so far, the progress on and beyond PoCs and pilots and the emergent role of service providers. The investments today are helping establish new norms for people, processes, and technology that will pave the way for “industrial scale” Watson in the future.

Posted in : cognitive, intelligent-automation


Leave a Reply

Your email address will not be published. Required fields are marked *

    Continue Reading