There are a lot of questions about how artificial intelligence supports knowledge management and how knowledge management supports AI programs. On the one hand, chatbots are a channel to knowledge, content and data and therefore needs to be correctly structured, tagged and curated to support a set of use cases. On the other, AI can be used to improve content access and retrieval through semantic processing.
In either case, fundamental information management principles still apply:
- the source of data and content needs to be reliable;
- the information accurate and suited to the application;
- a framework for organizing the information – whether embedded in the AI program or developed outside of it – needs to be applied.
While some programs can begin to make sense of messy and unstructured data and content, the data cannot be of poor quality. In the case of chatbots, semantic processing of user questions is applied to understand the user’s objective with a rich enough set of signals and cues to allow the bot to return a meaningful answer. Machine learning can, in some cases, derive dialog structures from large sets of chat log data, however this still needs to be cleansed and normalized by a knowledge engineer.
In this executive roundtable, our panel will explore foundational concepts of knowledge engineering that need to be considered in your AI project.