The enterprise or strategic vision for Watson (as with most tools and technologies) is based upon an evolutionary roadmap that begins with the initial evaluation of and experimenting with (Watson), to the predicting and assembling and finally (hopefully) the management and optimization of your organization's Watson use.
This initial phase will generally cover the discovery and evaluation that takes place with any new (to the organization) tool or technology. Testing should cover everything required to determine if the tool/technology will meet or exceed the organizations needs now and into the foreseeable future. Once a decision is made to move forward, this phase also includes the installation, configuration, and everything that may be involved in deploying the new tool or technology for use by the intended users of the organization.
From a Watson perspective, this phase may involve accessing the free version, before upgrading and subscribing to a premium version of Watson. In addition, Watson should be validated using actual data samplings, in an environment comparable to the environment that will eventually be used. The Watson documentation advises that output from Watson may differ somewhat when using certain operating systems: the results obtained in other operating environments may vary significantly.
Once you've accessed Watson for the first time, you'll begin exploring, predicting, and assembling. The effectiveness at which these events occur and the level of quality of those developed artifacts or assets (more on this topic later in this chapter) will be contingent upon the (hopefully escalating) level of experience (of Watson) that you possess. Characteristically, the exploration and use of Watson will be recurrent with each success.
Watson is a commanding instrument and, in most cases, does not necessitate extensive training to begin using. With human nature being what it is, once a reasonable output is realized and is validated as correct and valuable, it will be set aside for later reuse and, hopefully, pooled as a valuable knowledge object across the group or entire organization (either through a manual share or, if you have upgraded your Watson to a premium version, you can share collections more easily).
Although most organizations do eventually reach this phase, it is not unusual to see:
- Objects (everything from duplicated files, predictions, dashboards and stories) with similar or duplicate functionality
- Poor naming of objects (you've forgotten the details of the artifact so simply recreated what you need now)
- Objects not shared (but kept in collections indefinitely)
- Objects that are obsolete or do not yield expected results
At some point, usually while (or after a certain amount of) artifacts are developed, the process will mature. Process maturity starts with beginning to devote time and perhaps dedicating resources to organize, manage and optimize what has been developed and saved (or in other words, perform organizational knowledge management).
You should consider forming a governing committee to support the identification and management of your Watson knowledge within your organization to:
- Record and appraise Watson output and insights
- Establish suitable naming standards and styles
- Establish suitable process and development standards
- Create, implement, and impose a strict validation strategy
- Continually develop a vision for Watson within your organization and the worldwide Watson community
As I've mentioned before, as a part of taking a strategic or enterprise approach to Watson, you'll need to develop a vision. This will ensure that your organization is able to leverage Watson to achieve the highest rate of return today and over time. This is not simply uploading data files and creating predictions, but it involves everything (and more) that we've discussed in this chapter best practices and continued improvement of Watson skills.