For the modern consumer, “smart technology” is regularly influencing their decision-making. A TV fan just finished binge watching an Emmy-nominated drama, and is suddenly presented with a group of show that may pique their interest. A person gets an email from their favorite online shopping site promoting items, almost all of which are of interest. How do they know? Cognitive computing is how.
While there is no HAL or SkyNet lurking behind the scenes, cognitive computing itself a type of artificial intelligence. Cognitive computing aims to provide an experience to the user that mimics that of an interaction with another human. Think of it like this...if you ask a friend, spouse or family member for a recommendation on a good book or restaurant to try, they’ll provide their feedback based on understanding of the context of your question and what they know about you. In short, this is what cognitive technologies are looking to accomplish.
The goal is to provide contextual, personalized interactions to the user through ingestion of unstructured data and by applying machine learning to how users process this data. Applying this to procurement, cognitive buying means that the buying experience will vary from user to user so that it’s uniquely suited for that particular person. Over time, this experience will improve as additional data is ingested and analyzed.
For our clients’ users, that means the more they interact with the site, the more data about their interest is captured and used to create a one-to-one user experience. This results in a more consumer-like, intuitive shopping experience for the user which means less maverick spend, more on-catalog spend, less time spent on requisitioning and greatly increased user satisfaction.
The key is data. The system is only as smart as the data it has. Using the TV example, watching live TV provides no information to an online viewing tool, creating an incomplete picture due to the gap in data. Complete ownership of data and its history is critical to cognitive buying. So any solution that requires a user to go to an external suppliers’ website via “punchout” will not only be unable to do cognitive buying or “guided buying” so should be looked at critically.
In comparison, BuyerQuest owns the entire buyer journey and so has been able to apply cognitive buying to several aspects of our technology.
- Contextual search: Content groups and robust searching allows for contextual searching where results vary for different users. For example in a large media company someone from the sports division who searches for “Camera” will see tv cameras and parts in the results, but for someone involved in on-site entertainment will see security cameras, while another division may see disposable cameras. This saves time for each group.
- Behavioral analytics: Taking the best from online retailers, we track shoppers like an online retailer would by understanding their behavioral attributes. For example, when combining behavior with Net Promoter Score (NPS), the system can determine which behaviors line up with a positive experience and adjust accordingly. For example, a large education services company tracks NPS at checkout in order to understand how users feel about their overall experience. This data can be aligned with the actual shopping experience to see what occurred to generate to that score. This creates the opportunity to apply machine learning to mitigate potential negative experiences before they happen.
- Workflows: The enablement of personalized workflows based on granular conditions with exception based handling means that no two people may experience the same process conditions. As part of cognitive buying, actual workflow processes are unknown up until the conditions are set. This results in a personalized process for every user. As mentioned in a previous post, all customers use some form of workflow. Cognitive buying allows for an unlimited number of permutations
- Supplier Insights via IBM Watson: Parsing data from around the world to display only relevant global supplier is displayed and available for viewing.
As mentioned at the beginning, successful cognitive buying requires complete data. Because BuyerQuest eliminates Punch-Outs, the system retains all of the data associated with the procurement process, powering an intelligent, intuitive buying experience.