Your guide to decoding cognitive business
Early adopters of artificial intelligence share their insights
Cognitive and AI capabilities promise to transform organizations, but adopting this new technology can seem daunting — even for sophisticated IT users.
However, according to a new IBM report on decoding cognitive business, cognitive pioneers say adoption doesn’t have to be overwhelming. These organizations already apply cognitive technology to accomplish targeted business goals — from evolving customer acquisition to personalizing health programs to reinventing risk management. IBM research shows that 65 percent of early adopters believe cognitive adoption is very important to their strategy and success, and more than half regard it as a must-have to remain competitive. Many already see benefits from their cognitive initiatives:
How do pioneers seize the cognitive advantage? The experiences of companies like Swiss Re, Macy’s, Influential, RSG Media and The North Face show a variety of approaches to get started with cognitive projects and some best practices for those taking the cognitive leap. Adoption varies from bold, enterprise-wide implementations to more gradual deployments. Some organizations create custom platforms, while others use composable APIs.
They also leverage a range of capabilities, from machine learning to natural language processing, to unlock value from data sources, both structured and unstructured. Regardless of their approach, there are some commonalities that early adopters prescribe to.
Choose your on-ramp Many cognitive pioneers start with a pilot project, focusing on a business issue to innovate around that affects a critical mass of customers. These organizations have also learned to embrace a test-and-learn mindset and continually iterate and refine models and algorithms. “A lot of companies over-analyze what they should be doing,” said Paul Van Der Hulst, founder and director of Jibes Software, which develops and implements cognitive solutions. “And that means they want to exclude all risk. It doesn’t work that way. It’s better to start small and scale from there. There is no universal template for success, but focus and persistence is a proven formula.” A good use case should be scalable and have the data to support it. If it’s too discrete, the investment will benefit only a narrow slice of the business.
Advance your data strategy In the age where data is king, cognitive pioneers know it’s essential to be clear on what data sources are needed, where that data is being created, and how it’s being managed. “The management and curation of data is probably one of the biggest things that clients underestimate,” said Holt Adams, Executive IT Architect at IBM. An organization needs enough relevant, high-quality data to allow the cognitive platform to generate statistically valid hypotheses. The business application can’t be so futuristic, for instance, that the data doesn’t yet exist. “Being able to build a data model that allows companies to take data that is in different formats and structures and reuse that information in a meaningful way is enormously powerful,” says Adams.
Team for success Large-scale transformation projects may need extensive training to build a corpus of knowledge for the cognition process to be effective. But even for small scale projects to take flight, cognitive pioneers say teaming is essential to tap a variety of skill sets to support the effort. While it can be tempting to hand over the leadership reigns to IT, the most successful cognitive initiatives are cross-disciplinary projects with heavy business involvement. Working teams now include employees from lines of business, IT, research and a whole crop of data scientists and analysts. Cognitive pioneers also rely on an ecosystem of experts from IT consulting companies to industry analysts to help address and augment critical skills.
Originally published on the IBM Watson blog on November 29, 2016.