When a support case is opened, nothing is more frustrating for the agent than not understanding the context of the case. Agents often struggle with too much information (and sometimes too little) to really understand the customer problem. Trying to decipher that—and swiveling between applications to find what they think is the most relevant information is stressful—and can lead to agent burnout. 

In our last blog, we discussed how insight panels are a terrific remedy for both these issues, as well as six additional ways insight panels can be useful throughout your company. In this article, we’ll look at how you would equip your agents with the most relevant information—all in one interface. 

Here are some high-level best practices on how insight panels can be embedded directly into your customer service application (like Salesforce, Zendesk, ServiceNow).  

Elements of an Insight Panel

First, let’s define the elements of an insight panel:

  1. Unified case-solving content 
  2. Contextual signals
  3. Machine learning
  4. User profile service

Unified Case-solving Content

Make a list of all those applications that agents are diving in and out of—and think about what content within those apps is the most important. Some examples of content types that agents might need include: Software bug tracker, product documentation, support videos, knowledge articles, internal wiki, discussion forums, Slack, and more. 

Understanding customer wants means understanding the context of where they have been searching. When a user makes a query and a context is specified, the matching relevant results with the highest activity history in this context are promoted. Same goes for automatic results in the insight panel… results that have been helpful in a similar context are promoted, automatically.

Contextual Signals

Contextual signals are a form of behavioral data that tells us what individuals have searched for, clicked on, not clicked on. You can collect signals on customers as well as agents. For customers, we can see the journey they took right up to activating a support/service request ticket.

The support (or service) request form is an important web-based support tool that is frequently overlooked. It is usually found under headers such as Contact Us, Contact Support, Submit a Ticket, or Create a Case. 

Often, it is a page with several fields for the frustrated and time-starved customer to input all their contact details, answer some questions, describe their issue, click submit, and begin the excruciating wait for help.

But content request forms don’t have to represent a black hole. We recently looked at which support ticket fields can aid successful case deflection, as well as chatbots and virtual agents. 

Machine Learning

Some of the more advanced features in search today are to offer query suggestions (autocomplete, typeahead) and automatic relevance tuning (re-ranking). In the natural language space, an application extracts terms based on linguistic rules and word frequency.

In addition to traditional natural language processing (NLP) methods, including intelligent term detection (ITD). Coveo Machine Learning (Coveo ML) leverages vocabulary previously employed by end-users to search for content. This provides a more accurate picture of what’s contextually relevant to the current end-user.

In the support space, people type in very long queries—too many to be keyword-driven. Instead, models learn the expressions, leveraging contextual signals, to refine the queries. 

User Profile Service

The user profile service analyzes behavioral data to better understand user interests, preferences, and intents. It also automatically stitches together sessions of the same user across all devices. It’s even used by the User Actions feature of the Insight Panel to view customer’s activity across all search interfaces (IPX, Community, Website, Support Portal, etc.) and is used by ML models to add additional personalization to the output of the models: better query suggestions, better-promoted documents, etc. 

As the models learn, both agents and customers benefit! 

It’s really important that the vendor you choose offers both historic and live (in-session) user profile services. 

Two different users will see two different query completions depending on their User Profiles.

Coveo provides all of these services in its Relevance Platform and can natively embed search, recommendation, and personalization results within the context of leading support software like Salesforce, ServiceNow, and Zendesk. 

Understanding Customer Case Context

As noted, a key frustration is not understanding customer context. Typically an agent receives a call or a form with a request for assistance. Without understanding the true intent of the customer, a search through systems can only return general results. It falls on the agents to affirm with the customer which results are relevant. 

This manual going back and forth with the customer is both time-intensive—and frustrating to all parties.  

Instead, the agent can benefit from having an insight panel that has captured the customer queries—and the customer’s actual intent. When the customer submits a ticket, all the details that were submitted become the context for recommendations for the agent. The agent is literally picking up right where the customer left off.

Supporting Agent Workflow and Interface Controls

Besides understanding customer intent, insight panels are an elegant solution to the swivel chair problem. By building the insight panel directly into the support end users’ workflow, you are able to reduce navigating between windows, reducing agent errors—and stress. 

Coveo’s out-of-the box solution lets you configure the insight panel with tabs and filters important for your customers and agents. This allows you to group content together in ways that are logical for your users. 

Create simple and clear tab names to help users know what to find on each tab and how to leverage the content. Below is an example of how they can be used. 

Consider this tab configuration:

The tabs are labeled as types of content from their respective content sources. This can be as granular or broad as you like. It’s a way for the administrator to choose how information can be accessed—either unified (ALL) or by individual collections. 

Since the customer had to open a case, we know that this content did not solve their issue. Knowing this helps tune the self-service relevance. Additionally, it allows the agent to understand the steps the customer has already taken, providing the agent with the ability to point the customer to self-service options they might have missed.

The User Actions button, shown above, allows the agent to view any content that the customer may have viewed online before creating the case.

The design can include result actions. For example, the panel enables agents to attach items to cases, email links, post items to a feed, or copy the document link. These quick actions allow the agent to maneuver quickly through the case and avoid copy/paste and cross-window navigation.

The key to creating an effective insight panel is understanding the agent’s workflow. 

Every client implementation is different. Coveo provides the tools to configure the agent panel to assist the workflow, speeding adoption. Rather than changing the workflow to adhere to an experience—adapt the experience to the agent’s workflow. 

Coveo doesn’t want to deliver just a “search experience”; we want to build search that helps agents stay in the flow of work.

Dig Deeper

Wondering how you can improve your customer service KPIs? Learn from what’s been successful for others—in our How to Make Customer Service KPIs Soar ebook.

Share this story:

About Neil Kostecki

Neil is Director of Product Management at Coveo and has over a decade of experience across key Canadian tech companies, with a deep understanding of Service solutions and concepts. Since joining Coveo in 2017, his passion for content, search, and personalization has been a key driver in the evolution of our Service products. As an avid mountain biker, you can catch him living the #CoveoLife shredding trails at some of Ontario’s top spots like Hydrocut in Waterloo (where he previously worked at BlackBerry managing their KCS practice), or Kelso in Milton.

Read more from this author

About Diane Burley

Diane Burley is an expert on what it takes to stand out in the digital age. She showcases her expertise when writing about the necessity of AI, NLP, and intelligent search in DX projects. She cut her teeth as a reporter, is a long-time technologist, and knows how to tell a story in a gripping way

Read more from this author