“Well, it’s interesting. Whenever I’m asked what’s new in KM, I often respond by saying, ‘Not much,’” Garfield quipped, before following up quickly with, “OK, and obviously AI has become a hot thing now, and it’s widespread in its presence, and everyone’s fascinated.” And although the current run of AI only started 2 years ago, he sees it as “overexposed,” but thinks, “We’re underprepared for the practical use cases.”
Garfield cautions practitioners about “going after AI with the expectation that it should be able to do almost anything.” Not everyone is fully prepared for how to actually use it effectively.
Zach Wahl, the CEO of Enterprise Knowledge, recently addressed the SIKM Leaders Community to talk about what AI can do for KM and what KM can do for AI. One of the things KM can do for AI is to empower learning communities to help people learn how AI applies to them as well as how to navigate through changes generated by it.
Large Language Models
“Ideas like taxonomies have been around, as has AI. What’s new,” Garfield observes, “is the large language model tools that have a lot more power than we’re used to. Harnessing that becomes a challenge.” He suggests that concepts such as design thinking, usability, user experience, and agile methodologies are good topics to apply to AI within organizations. Those are not typical KM topics, but Garfield sees the need to adopt them as crucial in order to traverse the intersections between AI and KM.
“KM is a big umbrella and lots of things fall under it. All those things can be helpful. In my workshop, I show them 100 KM specialties that I’ve defined. I say it’s impossible for anyone to master more than a few of these.
“We have to use other people to handle those we don’t master. We need to impress on attendees who are starting out on their KM journey that they need to pick something—and that something may not be AI. You have to be focused. You can’t try to do it all, but there’s a lot of help available. Conferences like KMWorld allow you to seek out and take advantage of the skills of others.”
Part of that support will come from AI. Garfield sees some community tasks that AI will likely be able to take over soon, such as tagging content, managing threads, renaming and reorganizing content, monitoring discussions, and perhaps even helping keep discussions on track within threads. These moderation functions can remove some of the mundane work from community management, allowing participants to focus on dialogue and ideas rather than community management tasks. It could also summarize monthly calls and post them. Garfield would love to see someone prepare the SIKM Leaders Community’s 19-year-old threads to be consumed by AI, making all of that knowledge more accessible to SIKM members.
High Expectations
Garfield cautioned his workshop attendees about the sky-high expectations for AI. It will be hard. Management will expect KM departments to bring enterprise content into AI in the same way they see general content being brought into chat experiences. He suggests that like search before it, enterprise AI can’t perform as quickly or at the same scale as highly funded, focused internet services. Organizations are not yet prepared for the engineering and content preparation required to create an internal AI experience that is similar to external tools.
That said, Garfield senses that the “whole question of KM is about to change. When I start out the workshop, I show 4 different eras of KM. AI will define era number 4.”
After 30 years in this field, Garfield has seen tools come and go, but the fundamentals remain the same.
“We need to continue doing the right things,” Garfield concludes. “We can’t just discard essential practices that work in favor of a new tool, especially when that tool’s success is so dependent on all of those practices.”
Any organization that doesn’t understand KM’s role in AI will likely struggle to effectively integrate it into their enterprise systems and experiences. Garfield plans to deliver his KM 101 workshop again next year. Perhaps next year more senior leaders will attend so they can better understand not just what is possible, but what it takes to realize AI’s potential.