AI-powered knowledge intelligence: the next-gen of knowledge management

Optimizing for CX and agent engagement

Add bookmark

Kindra Cooper
12/12/2018

As with anything overhyped - the retail apocalypse, the gig economy or the Malthusian ceiling - our expectations overshoot reality.

It’s the same with AI, which many businesses perceive as a universal salve for problems like unmotivated agents, tedious manual workflows or overburdened phone support systems.

Implementing a knowledge management base or AI-powered call routing system should be a strategic decision centered around a specific business problem, says Sedarius Tekara Perrotta, chief product evangelist at knowledge sharing platform Shelf.

“You have to meet as an executive team and figure out the problems you want to solve and then go through a process of vetting and seeing how AI helps solve that problem,” Perrotta said in an online event with CCW Digital, ‘Grading AI Ahead of 2019.’

AI is a time commitment: what you should know before you buy

A knowledge management system is a time commitment not only to implement but to regularly maintain and revise knowledge entries so the information is accurate and up-to-date - far from a rote data entry task. Those entries then need to be optimized for searchability - meaning they should be indexed in a way that corresponds to the keywords agents use to search for things, like any search engine.

The final task is to establish clear metrics and objectives for the knowledge management system so you can determine whether it helps or hinders agents in live interactions with customers, and whether customers can easily find the answers they’re looking for through self-service portals.

Given the level of upkeep and optimization needed to truly leverage a knowledge base, the power of AI to root out inconsistencies and identify bird’s-eye-view trends from data is imperative. “AI can see the patterns of usage and see patterns of things going out of date and recommend improvements,” Perrotta explained.

According to a CCW Digital Market Study, The Future of the Contact Center in 2019, knowledge management represents the #2 strategic priority for 2019. However, the overhyping of AI combined with a lack of education on how to leverage the technology often leaves early adopters feeling shortchanged. Hence the need for Knowledge Management 2.0, or, as we like to call it, “Knowledge Intelligence.”

Why a uniform knowledge base is crucial

In many businesses, information is inconsistent. The marketing manager writes copy for the website, while the contact center manager inputs entries in the knowledge base; meanwhile, the PR team has its own version of the truth. Many organizations store information in disparate forms, ranging from the content repository such as a blog or customer support forum, the knowledge base of FAQs on the website or the agent-facing dashboard, and the information in self-service portals.

“Having three, four or more sources of truth makes it very difficult to execute a clear and consistent communication strategy with customers,” says Perrotta. Fragmented knowledge between systems forces employees to search numerous, often conflicting sources of information, especially if organizations store and manage data in different places. The chatbot, IVR, phone support team, live chat group and social media managers should have access to the same information.

If an agent delivers erroneous information to a customer because of an outdated knowledge management base, they start to distrust the system. It also creates what Perrotta calls “tribal knowledge,” where agents develop their own versions of the initial information based on trial and error and share it with a select few on their team, creating further fragmentation.

“If you have a lot of people who only have specialized knowledge, and they’re just talking amongst themselves that’s a real problem, because when they leave you have these big gaps to fill,” said Perrotta. “What a knowledge base does is it diffuses the knowledge of any particular person.”

AI-powered knowledge bases measure success

A knowledge intelligence system uses AI to handle mechanical tasks, such as identifying most frequently searched terms by agents and customers, flagging outdated or incorrect information, which knowledge entries or policies correlate to positive or negative customer experiences.

In terms of enhancing usability, AI assists with mechanical tasks like locating content, reconciling information, helping managers create, update and categorize knowledge entries. Finally, the AI provides a real-time dashboard that indicates the health of the knowledge management solution including how well agents use the platform, and which pieces of knowledge help or hurt the customer experience.

Once these important technicalities are handled by AI, “knowledge workers” can adopt more strategic roles such as setting the right metrics and reviewing performance data. Perrotta recommends that the knowledge management admin login once every two weeks for a quick sweep to ensure everything is up-to-date.

Your checklist for a good knowledge management system: