Semantic Layers Bring Answers to Problems KM Is Designed to Solve
The second answers the classic statement of “We don’t know what we don’t know” that many organizations continue to lament. With a semantic layer framework, an organization can actually spot where they lack explicit knowledge and information, or where people are asking questions for which explicit answers don’t exist. In these cases, the framework can help to automate filling those gaps by prompting experts to answer questions or tapping those with authority to provide coverage.
Lastly, the common challenge of many organizations continues to be harnessing tacit knowledge, maximizing knowledge capture and transfer. Too many organizations have invested heavily to solve this challenge, but ended up with processes that couldn’t scale or were too difficult to sustain. Semantic layers can help to automate these efforts where they were often most time-consuming, for instance in tagging content, restructuring it, and connecting it to other assets. When you combine these capabilities with the current slate of transcription and note-taking solutions, you can suddenly scale knowledge capture for the enterprise in a way that was never before possible.
Of course, all of these solutions still need a human involved, but they enable those humans to be involved where it is most critical and valuable, rather than doing repetitive tasks or “shuffling papers.”
In the book, you go into extensive detail about designing and implementing a semantic layer. What role do you see for KM’ers in this activity? Should they design and implement or should they oversee, manage, and contribute to the process? How about actually introducing the notion of semantic layers to their IT counterparts?
Any and all of those roles make sense, depending on the skill set of the KM’er and where they sit within the organization. Those that have focused more on the tacit knowledge capture side of the business will find themselves helping to fill knowledge gaps to build a robust framework, those with a semantic background will be more involved with design, and those that traditionally had a more data- or content-centric role will like focus more on knowledge asset readiness. Regardless, all KM’ers should get smart on the topic in order to guide these initiatives within their organization, translating the business value of it for their business stakeholders, and potentially serving as the stewards of the framework as it matures.
AI and semantic layers go hand-in-hand. The various components of generative and agentic AI change at an astonishing pace. Given the interrelationship between the semantic layer framework and AI, what do you see coming in the future, both short and long term?
Things are, in fact, changing so rapidly that from the time we started the book to when we finished the first draft, AI had taken hold as most organizations’ top priority. We predicted that would happen, but not nearly as quickly as it did. In the near term, there will be many organizations having invested millions into AI and having yet to realize returns, or, worse yet, having failed to accomplish anything meaningful. These organizations will, and are already, waking up to the role a semantic layer framework can play in delivering highly contextualized, trustworthy, and understandable AI.
In the longer term, and by that we mean the next 2–3 years, we’re going to see some big mistakes with AI. As organizations unite all of these assets from different sources and get bolder with automation, there will be accidental releases of sensitive or proprietary information due to poor handling of security. We’ll also see AI accidentally surfacing old material, potentially with serious consequences in some industries. Of course, the semantic layer can help to address these risks as well, but it will take more than just good semantics.
Thanks, Joe, Lulit, and Zach, for writing this book and for all you do to contribute to the success of the KM community.