As International Business Machines (IBM -0.01%) adapts to the rise of cloud computing, the company is focusing on its strengths, building a strategy around high-value services instead of commodity infrastructure. One of these services is Watson, the cognitive computing system IBM originally built to parse and interpret natural language on the game show Jeopardy.

Watson is now available as a cloud service, offering more than a dozen different features for those developing applications, like text-to-speech and visual recognition, where the system can determine the contents of an image or video without any other information. Watson is a machine learning system, which means it needs to be trained to perform specific tasks by being fed huge amounts of data.

IBM has high hopes for Watson. The company has invested more than $1 billion in the Watson group, including $100 million dedicated to venture investments in companies building Watson-powered applications. CEO Virginia Rometty has said she hopes Watson will become a $10 billion business for IBM within 10 years.

While Watson is nowhere close to being that big today, the system is being used in a variety of interesting ways. Other companies, like Microsoft (MSFT -0.11%) and NVIDIA (NVDA -1.71%), have introduced products that use machine learning as well, so it's clear that IBM isn't alone in thinking that a system like Watson has real potential. Here are a few ways Watson is being used today, and what it means for IBM.

An IBM cookbook and a cloud-powered toy
In a couple of months, a cookbook titled Cognitive Cooking with Chef Watson will be released, the result of a partnership between IBM and the Institute of Culinary Education. All of the recipes in the cookbook, some of which are already available, were created with the help of Watson. Vast amounts of recipe and ingredient data were fed into the system, and Watson was able to analyze the relationships between ingredients and come up with novel combinations.

Source: Amazon

Resulting recipes include an Austrian Chocolate Burrito, a beef burrito featuring both dark chocolate and apricot, and a Vienamese-Apple-Kebab. In total, the cookbook will feature 65 original recipes created by Watson, and future applications of the system could include helping those with dietary restrictions.

Another interesting application of Watson is in the toy industry. Elemental Path, a company that entered and won IBM's Watson Mobile App Developer Challenge, is currently working to bring its Watson-powered CogniToys to market. CogniToys connect to Watson in the cloud, allowing children to interact by asking the toy questions and receiving answers. Watson can provide different content based on the age of the child, and parents can control the content through a web portal.

Source: Elemental Path.

Elemental Path is currently running a Kickstarter campaign for the toy, expected to start shipping in November, and the project has already more than tripled its goal. Whether these toys work as intended remains to be seen, but if successful, CogniToys could create a whole new category of toys, with IBM at the center.

The potential of machine learning
While cognitive cooking and smart toys are unlikely to move the needle for IBM, it's the long-term potential of machine learning that represents a big opportunity for the company. Already, Watson is being used in industries ranging from health care to finance, and with the service available in the cloud, any developer can build machine learning applications.

IBM's business is helping other businesses, and Watson offers tools meant to help businesses succeed. Watson can be used to analyze the risk of employees leaving the company, for example, allowing human resources to intervene and have a better chance at retaining top employees. In marketing, Watson can analyze campaigns and rank leads in order to maximize conversion rates.

These are just a few examples of what Watson can do, and IBM is actively expanding the capabilities of the system. IBM isn't alone, though, in recognizing the potential of machine learning. Microsoft recently launched its own machine learning service on its Azure cloud platform, allowing developers to create cloud-based applications, targeting the same types of use cases IBM is targeting with Watson. One example service provided by Microsoft allows companies to predict when a customer is likely to end a subscription, allowing the company to attempt to retain them.

NVIDIA is also using a machine learning system, although in a far more specialized manner. The graphics company recently unveiled its in-car computer system, which uses cameras and machine learning to detect specific objects, like stop signs and pedestrians. Semi-autonomous and fully autonomous vehicles will need to be able to accurately evaluate the environment, and machine learning makes this possible.

Machine learning is one area where IBM is ahead of the curve, and it seems very likely that machine learning will become an increasingly important technology going forward. Microsoft's challenge to Watson, while introducing competition, validates IBM's decision to focus heavily on the technology. As IBM plods through its current transition, machine learning and Watson represent a major growth opportunity for the century-old company.