The new knowledge managers are consultants, not librarians

Summary

The days of manually curating and organising company knowledge will soon be behind us. Knowledge managers can’t operate as librarians anymore. They must, instead, elevate their AI literacy, implement an AI-first KM stack and deploy themselves as consultants to implement knowledge-enabled workflows.

A few months back, I wrote about the AI-first KM manifesto. The premise of the manifesto is that AI platforms do much of the sense-making that we’d expect manual content curators to do in the past. So, while there may still be value in structured, elegant, browseable repositories of information, every generation of LLMs makes it easier for us to make sense of unstructured, freeform knowledge artefacts.

One thing led to another, and I ended up building a few examples of a knowledge stack that derives from the values and principles of our manifesto. The stack has four layers:

1. The productivity layer. People create knowledge artefacts in this layer as part of their work. Tools such as documents, spreadsheets, collaborative whiteboards, and meeting transcripts reside in this layer.

2. The navigation layer. This layer adopts a decentralised information architecture. Teams, not “knowledge managers”, organise their information because the platform allows such flexibility and subject matter experts know how best to navigate their content and reorganise it when necessary. A platform like Confluence enables teams to own their spaces and utilise consistent patterns and templates across the company.

3. The sense-making layer. Perplexity and other AI-enabled search tools help us find answers to our questions by crawling the internet. Similarly, AI-powered search and RAG-powered assistants from companies like Glean allow us to make sense of the unstructured and structured knowledge from the bottom two layers.

4. The generative + agentic layer. To me, this is the promised land of AI-first KM. “Knowledge management” is not the end-all. The purpose of any shared knowledge is to support new, more informed, more intelligent actions. As we enter the agentic age of AI, the generative and agentic layer of a modern knowledge stack powers knowledge-enabled workflows. 

An example of a multi-layer KM stack

A new way to think about knowledge management

If you examine the above stack, you’ll notice that its decentralised design can work without much intervention. A knowledge management team’s job isn’t just to create or curate content anymore. Companies should expect erstwhile knowledge managers and curators to put the right systems in place and then get out of the way.

However, extracting value from AI takes time, as BCG has previously reported. Their analysis was telling.

“Companies face numerous challenges when implementing AI initiatives, with around 70% stemming from people- and process-related issues, 20% attributed to technology problems, and only 10% involving AI algorithms — despite the latter often consuming a disproportionate amount of organisational time and resources. Too many lagging companies make the mistake of prioritising the technical issues over the human ones.”

AI tools have surpassed a threshold of capabilities. You could argue that the features of many AI platforms exceed most companies’ ability to adopt them. This mismatch of capabilities and organisational appetite presents a splendid opportunity for individuals who have otherwise engaged in traditional knowledge management work. They can refashion their roles as business consultants. Think of them as consultants who help stakeholders within the company leverage their knowledge in AI-powered workflows.

Here are 10 examples to consider.

1. How do developers use a company's coding style guide in their IDEs?

2. How can salespeople use enriched customer information in their CRMs and internal research reports to customise and automate their outbound campaigns?

3. How can support teams leverage past issues and product documentation to resolve customer issues at speed?

4. How can engineering managers leverage activity on code repositories, Confluence, Jira and GitHub to build automated progress reports that’d otherwise take them hours to craft?

5. How can team members use AI to embrace artefacts and documentation, reduce endless hours of meetings and free up time for creative work?

6. How can HR onboard new employees by providing a friendly chatbot that patiently teaches them everything they need to know about their new job? No stupid questions!

7. How can B2B sellers leverage their company’s product or offering documentation to build rich and compelling proposals for their clients? 

8. How can infrastructure teams leverage system logs and past incident reports to predict and proactively resolve potential outages?

9. How can product designers leverage historical design feedback and analytics data to recommend UX improvements directly within their design workflows?

10. How can technologists integrate their documentation repositories directly into their IDEs to surface relevant architectural decisions or code snippets during development?

You’ll realise that there are countless ways to leverage team, company and publicly available knowledge to power more efficient workflows. Companies that implement more such workflows will reap productivity gains that’ll likely give them an edge over their competition. Traditional, librarian-like knowledge management (KM) is not yet defunct, but I expect it to account for only 10-20% of a modern enterprise’s knowledge strategy. That leaves significant capacity for KM professionals to wear the consulting hat. 

From librarians to consultants

I have a call to action for everyone in a traditional knowledge management (KM) role. Consider how you can refashion your role to be more AI-first. If not for your employers, for your career growth. I have a few suggestions to help you get started.

• Enhance your AI literacy. You needn’t be an expert at everything, but you must know your company’s AI toolset in and out. Subscribe to product updates for these tools. Learn how to push each platform to its limits. 

• Understand your company’s business as deeply as you can. Identify pain points worth addressing. Implement standard problem-solving tools, such as fishbones and process maps, to determine where a knowledge-enabled workflow can drive better outcomes.

• Find a willing stakeholder to partner with and help them elevate their productivity with a knowledge-enabled workflow. 

• Transfer your learnings to other teams, have them support each other and build a network of people that help each other implement new workflows. Tell stories and share your successes. Nothing’s more contagious than success, is there?

Enabling knowledge-powered workflows as a consultant

Think of the above steps not as a one-to-done sequence, but as a loop. The more you succeed, the more demand you’ll see for your consulting services, the more you’ll have to elevate your skills and so on! Isn’t that a place we’d all love to be in, though?