Zendesk on Zendesk: How we do KCS
Zendesk on Zendesk is a discussion about a specific topic and how the Zendesk Support team uses Zendesk products. Each session is hosted by a member of our Support team.
In this Zendesk on Zendesk, Ricardo, our KCS Specialist, explains how our KCS implementation matured from the early day's program described in Zendesk on Zendesk: How we use the Knowledge Capture app.
See all of the Zendesk on Zendesk series discussions.
**KCS® is a registered service mark of the Consortium for Service Innovation™.
Four practices for everyone
For the success of the program, we had to make things simple. Zendesk Guide offers its own tool that allows agents to easily input content into tickets and to easily capture knowledge from tickets into knowledge base articles. Our agents use the Knowledge Capture app and we encourage them to do any of the below tasks in all of their asynchronous tickets.
We train our agents to search throughout the knowledge base before replying to customers. Public documentation in Zendesk’s Help Center doesn’t require authentication to be accessed and because of that, all articles are indexed by Google. Agents can either use the app search bar or they can simply search in Google.
Reuse in our KCS implementation is called linking. Reused content is a very important source of data for us. For instance, the Browse common questions sections you see in each category page of our HC (for example Support – Help) are built around the most linked articles in our interface.
Flag or fix
Whenever our agents see an issue in one of our articles, they can either flag the article or fix the article if they have the right permissions. We train agents to when they flag an article to let us know not only what is wrong in the content, but also how to fix the article.
Our agents can also create content. Depending on their role, they can either create the content themselves or they can suggest the content to be created on their behalf.
Because we don’t follow all the guidelines provided by the Consortium for Service Innovation, we call our KCS implementation Knowledge Sharing.
Something that is central to Knowledge Sharing (KS) is to reward good decision-making. We created a structure of roles that rewards agents with autonomy and gives them the motivation to participate in the program. Additionally, it helps agents who already have some interest in self-service having contact with that world.
Every new hire in the Zendesk Customer Support team is trained as a Contributor. Contributors are usually agents with general knowledge of Zendesk products who are at the forefront of customer support. As a KS Contributor, they are expected to search, link, and flag articles and also suggest new content to be created.
A Self-Publisher can do everything that a Contributor can plus fix articles that belong to their area of expertise and also create their own articles. Agents who progress into more technical roles are trained as Self-Publishers. We are also open to train Contributors into becoming Self-Publishers if the agent shows interest in writing their own content.
Knowledge Champion role
Knowledge Champions are super KS agents. Those are agents who have shown interest in self-service and want to get more exposure to it.
All agents who have a leadership role are trained to be a KS Coach. This role is intended for the Managers who oversee a team of agents. They should be able to coach their agents to make the best KS decisions, to promote the program in their teams, and celebrate any KS achievements.
The Knowledge Champions program
The Knowledge Champions program is one of the most important parts of our KCS implementation. They are the bridge between the Self-Service team and the Zendesk Customer Support team. At the moment, they are a major part of our content strategy and they enable us to move really quickly in KCS and beyond.
Our Knowledge Champions are support agents whose main job is to help customers but who also have two hours of their day scheduled to work on Knowledge Sharing tasks. There is an extensive list of things a Knowledge Champion does, but to make it short their tasks vary from reviewing articles created by their colleagues, creating their own articles, updating articles based on ticket flags, and in general being an ambassador of the Self-service team in their own teams.
See below some of the characteristics we look for when choosing a Knowledge Champion:
- Strong advocate with a deep knowledge of Zendesk
- Understanding of how self-service impacts customers and interest in KM
- Ability to work independently
- Strong writer
Due to time constraints, sometimes it is difficult to offer them different tasks but we do try to give them up-skill opportunities. We want them to become a Knowledge Specialist. Being a Knowledge Champion is also a great way to try out what the Self-Service team does. In fact, some of our hires were Knowledge Champions, myself included :).
How do we use Zendesk Guide features?
We use different features and tools to make sure knowledge is reused, modified and created. Find below two examples of procedures that show how we utilize our own tools.
As mentioned before, our agents use the Knowledge Capture app to do KCS. They can either suggest articles for creation or create the articles themselves.
Suggest content to be created
When a Contributor suggests new content to be created, they are actually flagging an existing internal article and telling us what the article should contain.
- The agent searches for the iKB using a keyword like ks_new (or choose your own label) in the KC app. The article has questions that work as the template of the article-to-be.
- The agent inputs their feedback into each question with what they think the article should be about. At this point, we encourage agents to copy and paste their responses from the ticket where the knowledge comes from.
- A ticket is created in our KCS queue and a Knowledge Champion creates the article on their behalf. The Contributor is set as the author of the article so they can have recognition for their suggestions.
Self-Publishers are trained to create content themselves. Creating an article is also very simple if they use the KC app.
- The agent opens the KC app and clicks the plus (+) button or Create new knowledge.
- They choose a template for their article. We have two templates: Question & Answer and Issue Symptoms & Resolution Steps.
- The agent writes the articles based on their ticket. We want them to not think about formatting or text structure. If there is context enough in the ticket, they should copy and paste the interaction from the ticket.
- The agent clicks Submit for review and the article is then reviewed by one of our Knowledge Champions.
We use Team Publishing in our workflow because our agents used to share feedback that they didn’t feel comfortable publishing content without being reviewed. Another reason to use this feature is that we rely on machine translation. Writing content that is machine translated has its own guidelines. Only Knowledge Champions are trained on our writing guidelines because we want Knowledge Sharing to happen as fast and as easy as possible for our agents.
The Team Publishing feature allows us to add a layer of review to any article that is created or updated by agents. Even though we don’t give the option for agents to make articles public-facing themselves, we have an internal SLO of 48 hours to make sure the knowledge shared is published as fast as possible.
Knowledge Champions are responsible for reviewing the content produced by Knowledge Sharing (KS).
Process to review content
- The Knowledge Champion (KC) opens Guide > Guide admin.
- Under Manage articles, in Lists, they select Awaiting review and open any article listed in the workflow.
- The KC reviews the article according to our writing guidelines. These agents are fully trained in our content standard. Once the article is reviewed, they Publish the article.
- Then, open Support and in a new ticket apply a macro that notifies the author of the article. The KCs write down summed up bullet points of the changes they made in the article for coaching reasons.
What have we learned
Recognize value, not just volume
If we want agents to keep sharing their knowledge, rewards and recognition need to be front and center. We analyze Knowledge Capture activity, but we also look at other indicators, such as CTR in search results, page views, successfully served up by our answer bot. This data helps us understand if the authors are being successful at intercepting tickets.
We showcase good decision-makers in our team and help them export their best practices. We have a monthly newsletter where we highlight an agent by interviewing them. We hope to motivate and inspire other agents to participate in the program.
Make it really really easy
The only way to have a successful KCS implementation is if we make it really easy for agents to share their knowledge. We solicit regular, bi-annual feedback on our processes and tools to understand how we can simplify.
Feed the bot
The content that our agents create is perfect for our bot to serve. KS articles are short and focused on single topics, which makes it digestible to a machine.
To that end, we share our Answer Bot metrics along with KCS metrics and try to reiterate that our contributors play a big role in the bot’s success.
P.S. It is very nice to see all the improvements towards KCS alignment you did during the last years! =)
Thank you Andrei Kamarouski I think our experience could be helpful to other organizations!