Why Knowledge Management Fails
When I was SVP of products at ServiceNow, I frequently engaged with customers who were frustrated that while they had built a robust knowledge management practice, it had limited impact to their self-service initiative. Upon digging in deeper, we identified these issues:
- Knowledge managers did not always know where to focus their efforts to maximize the impact to the service desk.
- Some knowledge articles did not meet the expectations of employees. However, convincing employees to click on a feedback button was hard to achieve, providing limited visibility into the effectiveness of existing content.
- The search technologies used by self-service portals would generally return several unrelated articles, requiring employees to have to read through many documents in the hope one of them would address their concern.
- Being creatures of habit, employees tend to leverage whatever communication channel is most comfortable for them when seeking help, in many cases bypassing access to knowledge.
This last point has been exacerbated by the hybrid workforce. Without the ability to shoulder tap a colleague for help, there is an average 35% increase in tickets on an already overwhelmed service desk. A study conducted by Gatepoint Research found that although we are in the second year of Covid, most IT leaders have not redesigned employee support for the hybrid world. As a result, 74% of onsite employees and 80% of remote employees are calling the service desk – bypassing self-service portals.
Is Knowledge Management Dead?
No. There is nothing wrong with knowledge management. The issue is how employees want to consume content. As consumers in our home lives, we have all become very accustomed to quickly finding answers to our questions through virtual assistants such as Alexa, Google Home, and Siri. And yet, at work employees looking for information face intranets, portals, and an endless number of knowledge articles spread across multiple sources. This problem is further compounded by the fact that every department (e.g., IT, HR, Payroll, Facilities) has its own support processes.
There is also a big difference between our consumer and work lives. At home, we are basically on our own to find our own answers (well… except my parents for whom I am their own personal technical concierge). At work, we all know there is a team of people in IT (or HR, Payroll, etc.) who will “gladly” parse through that dizzying amount of content and find us an answer.
Why Not Let Service Desk Agents Do the Work? It’s Their Job, Right?
The reality is that most service desks are facing 41% attrition, and a 102% increase in absenteeism, which has led to a 68% increase in ticket backlog. So, while the service desk is willing and able to handle all calls, the reality is those resources would be best utilized for issues that really do require humans. It turns out that well over 70% of calls into the service desk are easily handled via a virtual agent.
Further, as millennials enter the workforce, they do not want to interact with people. They want their 9 to 5 to feel like their 5 to 9. They expect finding answers to their questions to be simple and prefer to use virtual assistants – as long as they meet their expectations.
How to Maximize the ROI on Your Knowledge Management Efforts
Now that we have established that the crux of the problem is one of experience, maximizing the return on your knowledge management investment requires a rethink of how your employees consume your content.
In today’s hybrid world, virtual agents have become the de facto approach for helping employees get automated resolution to their issues and requests. However, there are multiple options in virtual agents with distinctly different approaches to the problem.
- Toolkits: Products such as IBM Watson, Microsoft’s Bot Framework, and ServiceNow’s Virtual Agent are considered toolkits. While capable products, the level of effort required to build language recognition into these products is significant and requires an ongoing investment with expensive resources most organizations do not have access to (e.g., data scientists, computational linguists).
- Black box systems: Products such as Moveworks crawl knowledge to learn answers to employee questions. Some of these products train themselves on available knowledge base articles, and therefore these virtual agents’ helpfulness are only as good as the content they are provided by the customer. As black box systems, customers have no way to understand how to impact matching or accuracy. Therefore, customers are on their own to figure out how to change the behavior of the product by updating or creating more knowledge articles in the hope that it provides better outcomes. Further, these platforms generally do not provide an ability to perform regression testing to understand the impact updated or new articles may have on other related topics.
- No code, open platforms: Products such as Espressive Barista come as a fully managed service and are pre-built to understand the language of employees across the enterprise. A managed service means Espressive is responsible for linguistics and accuracy. This enables customers to focus on end results (e.g., automated resolutions) while the platform provides complete transparency on how the underlying AI engine behaves. Customers can have confidence on why something is matching or not – no more “guessing”. As customers update or create more knowledge articles, with the push of a button they can also run a comprehensive regression test in a development environment to understand the overall impact across all possible interactions, and then push to production with high confidence.
Ensuring the knowledge management team is focused on building content that will have a material impact to self-service efforts, measured via a reduction in service desk tickets, requires an approach that provides analytics to help guide the team. This requires products that provide analytics for the knowledge team to better understand whether their existing content is hitting the mark with employees and where adjustments are required to increase effectiveness.
Employee Change Management 101
One lesson learned from all self-service portal projects that I’ve encountered is that employees are creatures of habit and will not change their behavior for the convenience of the service desk. Any approach that requires employees to do something different will lead to limited success. And who needs more of that?How employees leverage virtual support agents differs across product.
- Dedicated application: These virtual agents typically exist in a separate application and require employees to go to a dedicated application to get the help they need.
- Portal Bots: These virtual agents are accessible via a self-service portal and require employees to access the portal in order to maximize the ROI. The issue here is that historically, less than 14% of employees would ever visit a portal and recent Gatepoint research shows this has not changed.
- Collaboration tools: There are a number of virtual agent products on the market that can only work with collaboration tools. While adoption of tools, such as Slack and Microsoft Teams, have gained popularity through the pandemic, the recent Gatepoint Research revealed that only 30% of employees go to those tools for IT help.
- Omni-channel: The only way to maximize adoption is through an omni-channel approach. This means a solution can interact with employees across multiple channels for support – wherever employees choose to go for help. In addition to apps and collaboration tools, employees can access the virtual agent via email, phone, integrated via an IVR, on SharePoint and intranets.
Summary
Knowledge management is not dead. However, your approach to knowledge management is critical if your goal is employee self-service and ticket deflection. These are the key things to remember:
- Don’t rely on tools that were not designed for the specific needs of the hybrid world (such as portals, phone, and email). Instead, leverage a virtual agent to deliver automated consumer-like employee self-help to enable a 24/7 hybrid team.
- Select a virtual agent that provides analytics to provide an ability for the knowledge team to better understand whether their existing content is hitting the mark and where adjustments are required to increase effectiveness.
- Select a virtual agent that enables you to focus on automation versus worrying about linguistics and accuracy.
- Select a virtual agent that enables you to understand the impact of new or updated content could have through a simple to execute regression test, ensuring your efforts are providing better outcomes.
- Finally, to achieve the highest adoption levels, your virtual agent needs to meet employees where they go for help, rather than trying to force them to learn new behavior. So while you are not relying on tools like phone, email, and portals, your virtual agent should have an omni-channel approach and meet them there as well.
The good news is that with a strong virtual support agent platform, knowledge management has never been more important. Your employees want answers contained in knowledge articles, they just want the virtual agent to do the heavy lifting for them, giving them the answer to the question or resolving the issue.
Now the ball is in your court.