S4 Ep. 2 Why knowledge management is a powerful customer experience asset

Coty Smith (01:16): 

Hi, I’m Cody Smith, the Director of Digital Innovation, working with the digital and AI groups here at Genesys to help our product, our customers, and our business adapt and transform to the ongoing digital innovation taking place internally, as well as out in the market.  

Coty Smith (01:55): 

Knowledge management is a core aspect of being able to manage your digital content as a single source of truth. And it’s doing that in a way that is channel agnostic, can be used in multiple ways, is dynamic and can update on an iterative basis, because information changes rapidly. It’s typically not content that somebody creates and sticks in a corner and says, “Great. We’re knowledge-enabled.” Knowledge management is an ongoing process of that digital content. 

Coty Smith (03:05): 

That’s correct. In an ideal scenario, knowledge is managed centrally. So, not only can it support an omnichannel experience, but it also could even go into a channel-less experience — as we start thinking beyond single channel experiences and getting into experience as a service management, where knowledge is front and center for that. 

Coty Smith (03:33): 

Yes. We’ve all seen knowledge management in a number of different ways before. In a lot of cases, we think of FAQ pages and such. It might just be on a static page; it might be a wall of text that you’re going through. But knowledge management is beyond that as well. It’s being able to help with creating elements of specific articles and being able to make that relevant to the end-user experience. That end-user may be external — maybe it’s a customer — but it could also be a colleague, somebody internal who might be accessing that same sort of information from an intranet page or other types of repositories of information.  

And we look at knowledge as part of our use case and flow, being able to put that into a customer experience or digital experience is part of that, as well. Being able to help enable self-service from knowledge management is a core value of enabling and empowering customers and users to serve their own needs. That’s probably the most natural starting point, but we know that knowledge can also take place throughout the different points of the conversation as well.  

One other thing I’ve seen come up is, while knowledge can be utilized with AI tools and services, it isn’t necessarily AI-enabled and it doesn’t have to be. It can be used in that search-response type use case, but it can also be utilized in those more enriched experiences where automation and bots and knowledge-assisted agent services can not only help provide the right answer, but also help with efficiency.  

Coty Smith (06:13): 

One of the most common scenarios is with agent-assist capabilities — being able to use knowledge to assist that agent based on what the customer need is. Also, it’s being able to do that across multiple channels all at once and being able to use the same knowledge that the customer may have had access to through a support center site. This ties back to having a single source of truth, and helps agents understand what the customer is asking for and what the most relevant knowledge may be. That can help not only with being able to respond quickly, but also more accurately and consistently. And those are the elements that, in aggregate, help improve the experience that the overall organization is delivering to their entire customer base.  

Coty Smith (07:30): 

There’s no wrong answer because it’s all connected. The biggest driver tends to be efficiency. When we look at the perspective of a user of knowledge, they have more options now to engage with an organization, whether that be through phone messaging, social and so forth. And often it’s the users and the customers who are opting for that efficiency; they know what their need is, and they feel like if they have the right tools, they can address it.  

If we have more users and more customers who are self-serving, that then makes agents available for more challenging work. Even though it starts with the efficiency for the customer, when we say efficiency, we would rather think of it as empowerment.  

Coty Smith (09:53): 

Absolutely. We’ve always looked at a knowledge base as a garden. You can go and plant it, but you have to prune it, you have to keep it updated, otherwise it gets overgrown, or nothing grows at all. And that’s not the best scenario when you’re talking about empowerment and enabling users and customers. 

When we talk about knowledge management as a process, part of that process is quality control, constant iteration. Using knowledge management tools should enable you to see not only what questions are being asked and answered, but more importantly, what questions are being asked that are going unanswered. What is answered incorrectly or maybe in an outdated way?  

And it’s that type of iteration that allows for ongoing improvement — being able to understand what the exact needs are gives a fresh lens to what customers are looking for. 

Also, we’re at a time where we’re hearing a lot about generative AI starting to come into place. We’re looking at how generative can help identify and improve upon those knowledge-led experiences — not just in terms of understanding what those questions are, but also helping serve as a process or tool in that process for helping with generating the knowledge response itself. I wouldn’t say that these are things that you just hand knowledge management off to; generative technologies are not appropriate for that. But it is something that we’re continuing to look at as to how we can incorporate its as part of our tools and services for doing it in an accurate and human, empathetic way.  

Coty Smith (14:01): 

The big thing is not putting yourself into that seat of a user. We’re all users of knowledge tools. We’ve all interacted with these types of things. And when you think about how you yourself may use this, you need to consider what types of things you would need to know, and what things that have challenged me, as well. It might be not having detailed enough information or, say, sizing charts that aren’t available.  

The other aspect is continuing to think like a user after deployment. It goes back to that quality control and optimization effort. The opportunity is how we can help our customers better understand something, because they’re looking at it from a completely different perspective. Maybe someone has to reset a password in a secure way and the steps make sense to us, but not to them; that will result in user frustration and reaching out to the contact center.  

It hard to classify that as an issue in all cases. But it’s an opportunity to think about what may be the reason. Let’s say sales and support, for example, need to be on different knowledge platforms. Does consolidation help make an improvement in the employee knowledge experience? 

From the end-user perspective, you wouldn’t want somebody to see something different on the marketing website than what they would see on the support site.  

We also know that being able to manage knowledge from one single system has values and benefits, as well. Everyone knows how to use it. These are things that can be unilaterally available. We might not necessarily want our external customers to have access to certain information, but at the same time, it’s helpful as an employee to know what our customers are seeing. It goes towards empathy and being able to understand and share that perspective.  

Coty Smith (18:32): 

That’s exactly it. Everybody’s coming into their knowledge management journey from different perspectives. From the starting point it’s understanding, what are your needs for knowledge? Are they available today? Are you using anything within the organization and how can knowledge help support that? It’s being able to help with getting those initial elements in place first, then go and support an AI-enabled knowledge management experience.  

Also, understand what the impact is of having knowledge. Is it empowerment? Does it reduce contact volume?  

Coty Smith (19:42): 

One of the biggest things to think about — and I’ll tie this back to thinking like a user—is knowing that there are many ways to ask for the same thing. And being able to think about phrasing and how users are asking questions. Again, a common example: how do I reset my password; can’t log in; locked out; what’s the process for getting access to my account? All of these relate to the same intent. We need to think about how we might include different phrasings so our knowledge can more easily help and adapt using the customer’s own language. Using these technologies that understand language can help us do just that.  

Coty Smith (21:22): 

I would say absolute must haves include the ability to create a process and manage a process for knowledge iteration. That’s table stakes.   

We’re also continuing to look at how well knowledge has a “steel thread” through it. There may be reasons to modify portions of text based on where a customer is coming in from. We’ll use our password reset scenario. Somebody asks that password reset in web messaging. They may get a full body of text with links and so forth. But if somebody were to ask that through SMS where there’s a character limitation, that’s where being able to have variation support to create verbiage specific for those different channels helps keep that intent-and-response pair the same and helps manage the experience in a more seamless way for each channel. 

Coty Smith (22:49): 

One of the biggest things that we see is how generative technology plays a part in knowledge management. My recommendation would be to not rely on it as your entire process, but rather as a tool for helping to understand the different nuances and phrasing that come through. This all ties back to a concept we call supervised learning because a knowledge management team can review what’s coming in and keep a clean data set in place.  

If we handed this off to generative technologies, you end up not understanding why content is added to your knowledge base. With a lack of curation — or unsupervised learning — things come in as a bit of a black box. That supervised learning approach, whether it includes generative or not, ensures that each part of that knowledge management process has human curation.