A beginners guide to knowledge gaps

[RogueFader]

Hi all.

I say a beginners guide, I joined this team to manage problem, however I seem to have had knowledge thrust upon me with no guidance, so I'm learning as I go.

I've been tasked with trying to address knowledge gaps in the public knowledge base, and I was just wondering how people approached this, maybe in a STAR approach.

Just to get some baselines, I've extracted 6 months of incidents where the category is 'User guidance' and where the incident is not linked to an article, but to be fair, getting our teams to link to articles if/when they use them is another battle altogether.

I then put that info into Excel, into a table and then used copilot to look at themes of issues and what types of articles we would need.

I'm sure I'll need to move it on from there a bit, although I'm not 100% on what reports I'll need to run, which will also need to include Virtual Agent, then scan all that data, and tie it all in together to address that knowledge gap backlog.

Any steer on this issue greatly appreciated.

Thanks

[DavidBReynolds]

The method I use--I'm the primary author of a knowledge base all employees can access--is looking at tickets whenever possible and then mentally filtering them to determine in an article can help. Articles apply to maybe 1 out of 10 - 20 tickets. In those cases, I search that knowledge. When I find an article matching the user's comments, I attach it to the ticket. When I don't, I follow the ticket. I'll either wait for it to get resolved or research it and get something up quickly.

We also use Moveworks Analytics, which is now part of ServiceNow. In this case, I'm looking at Topics Not Served to see what people are looking for then either update an existing article or write a new one. 

[Leri Andrews]

Hi

Train the agents to use the ‘knowledge gap’ flagging ability when they are working on a case and to attach knowledge to resolutions using the button rather than sending an email so that you can measure knowledge ‘use’. 

NowAssist (if you have it) can be used, but if you have an internal ChatGPT then you can ask it to review your anonymous case data and suggest themes. 

[David Kay]

Knowledge-Centered Success (KCS) is designed to solve exactly this issue.  As part of handling each customer inquiry in which knowledge is relevant (which is most inquiries in many environments), the person helping the customer searches the knowledge base.  If they find the answer, they attach or reuse it.  If they don't find the answer, well, there's a knowledge gap, and they structure their customer interaction to create a new article in a very streamlined format.  (AI can help.)  If they find content that's relevant but needs to be updated, they update it or flag it for review. 

There are lots of moving pieces that make this all work and assure quality, but this is the big idea, and filling gaps is just an intrinsic part of the process.

Of course, this assumes you have a single KB with a subset exposed to customers, but that's what you want anyhow.  It's tough enough to maintain one knowledge base: why have two separate ones for internal and external users?

[RogueFader]

Hi David.

So we have one knowledge base which is used by the technical teams for troubleshooting and incident handling. The other knowledge base consists entirely of guidance and easy fixes, Tier 0 support, that end users can simply do themselves.

The issue I've had is that when analysts search for an article, they almost never flag it as a gap, and the same issue of having them attach the article where they used it successfully to resolve an issue.