Cognitive computing: A diverse and fast-growing market
One of the keys to providing a quality interaction is to have specific and good-quality information available for each individual. “It is much better for the assistant to create a personalized engagement by thanking the person for attending a particular booth at a particular conference, or mention a specific white paper, than to say something generic,” Terry pointed out. “In the absence of that, we can default to a more general comment. Consistently low- quality data does pose a problem, though. “In those cases, the organization should develop some approaches for improving data quality, which will in turn improve the performance of the assistant.”
Given the widespread goal of automating more responses for customer interactions, it would seem likely that companies would be organizing their information for accessibility. “Unfortunately that hasn’t been the case in my experience to date, especially with semi-structured and unstructured information,” commented IDC’s Schubmehl. “It’s something that we keep emphasizing, but in my experience, most organizations do not have a comprehensive information architecture strategy. I think the emergence of chief data officers is helping with this, but we’re also in early days on that as well.”
Complex data analysis yields environmental insights
It would seem to be a commonsense conclusion that the natural beauty of the Caribbean would be considered to have economic value, but the data to justify and quantify that conclusion can be hard to find. Moreover, analyzing ecosystems and their interactions with the economy is a complex process. One recent such study conducted by The Nature Conservancy, in partnership with Microsoft and JetBlue, went a long way toward accomplishing the goal of clearly demonstrating the economic value of one ecosystem in this region: coral reefs.
Many threats to coral reefs have been documented, including rising seawater temperatures, physical destruction, and the impact of various forms of pollution, such as sediment deposits from urban development and storm water runoff, excessive nutrients from agricultural fertilizer, and pathogens from untreated sewage. The result is that a significant number of coral reefs have been lost, and many are threatened.
Several years ago, The Nature Conservancy received an in-kind cloud resource grant that enabled it to use Microsoft’s Azure cloud services, which provided access to Azure’s cognitive services API. The organization began working with Microsoft’s AI for Earth program to explore ways in which these new resources could be applied. The Nature Conservancy used resources such as cloud-based computing, AI, machine learning, and data visualization in its research. In addition, it partnered with Esri, the leading provider of geospatial software.
In this recent study, The Nature Conservancy used Microsoft’s machine learning capability to analyze social media content, including more than 86,000 images and 6.7 million texts to accurately identify reef-adjacent activities such as sailing, diving, and snorkeling among other beach activities. The image data containing geolocation information, which came from Flickr, was analyzed to determine density of photos taken in reef areas. This concentration of photos served as a proxy for tourism destinations, and once correlated with a commercial database of global accommodations, The Nature Conservancy was able to model the economic value of coral reef areas. Related studies also leveraged the cognitive services API to identify coral reefs themselves so users can examine both “on-reef tourism” and “adjacent-reef tourism” values.