Ada Support

The ultimate guide to building a knowledge base for customer service

Arjun Ruparelia
SaaS Writer and Content Marketer
Customer Service | 26 min read

Modern customers want to self-serve. Studies show that prefer to find solutions and answers themselves instead of speaking to a support agent. 91% of respondents specifically said they’d use an online knowledge base when available and tailored to their needs. That’s why most of your competitors have a knowledge base.

But how can your knowledge base deliver a customer experience that stands out?

In this guide, we discuss everything you need to build a customer service knowledge base that exceeds a modern customer’s expectations. We dive deep into how a knowledge base works, how it helps your business improve customer experience, and how AI can augment its power.

What is a knowledge base?

A knowledge base is an information library that enables customers to find information about a topic, product, or service.

Here’s what a knowledge base typically looks like:

A knowledge base includes content in various formats, such as articles, videos, and FAQs. So when your customer doesn’t know how to change their password, they don’t have to rely on support. They can read an article or view a video that illustrates the process.

Types of knowledge bases

A knowledge base can be internal or external. An internal and external knowledge base are conceptually the same but differ on two fronts: accessibility and content.

How to create a knowledge base for customer service

Some companies like to create one resource at a time. But if you have a need for speed, you can create all knowledge base resources at once.

The latter approach can be overwhelming. Your team will need to decide the content to cover, create text and video content, design the knowledge base, and organize the resources. But a well-defined process and software can make things easier.

Let’s walk through the step-by-step process to create a customer service knowledge base.

Step 1: Define your audience

The audience for an external knowledge base is your customers. You’ll need to include information about your product — tutorials, troubleshooting guides, company policies, and FAQs.

Internal knowledge bases are built for employees. An internal knowledge base includes all the information in the external knowledge base, plus other resources like training modules, compliance information, and onboarding materials. Your focus when creating an internal knowledge base for customer service should be to make finding information and relaying it to customers easier for support agents and AI-powered automation tools.

Step 2: Research and prioritize your subject matter

Thorough research is vital to prioritizing information in your knowledge base. You’ve probably got a clear understanding of your customer, but we recommend you go even deeper; thorough research helps provide relevant information.

If you skip this step, you won’t know what information to prioritize — and that has consequences, like losing customer loyalty and poor ROI on your knowledge base.

To find content topics for your knowledge base and ensure you’re sourcing information customers truly need, we recommend

  • Go through support tickets: Identify the most common issues your customers have by reviewing customer conversations with support and include the solutions in the knowledge base.
  • Review user journey: Map out a typical user’s journey to identify what questions they might have. Anticipate their information needs and address those needs when creating knowledge base resources.
  • Analyze use cases: Review your product’s use cases and identify situations where your customers might need assistance. For example, a user who wants to use a task management SaaS to set up reminders for monthly payments might need help setting up recurring reminders and configuring notification channels.
  • Seek feedback: Customer feedback gives you insights about your customers’ most common pain points. Provide solutions to these issues in your knowledge base. If too many customers are frustrated about having to spend too much time receiving a refund, create an article outlining the process and link to relevant sections of your refund policy.
  • Request cross-functional input: Ask employees in various departments to provide input on topics to include. For example, you can ask the customer service team about common issues they hear from customers and the product team about complex use cases and features to cover in the knowledge base.

Step 3: Organize content

Once you’ve put in the research time, you can start organizing the information. Information architecture requires careful thought, but here are some guidelines to organize your content:

  • Create categories and subcategories to organize content logically, but make sure the categories don’t overlap. When your content is logically categorized, humans and AI are less likely to look for information in the wrong category and waste time on irrelevant information.
  • Use relevant and descriptive signposts. Make sure the information under a signpost is highly relevant so users and AI never have trouble finding information. Put verbs closer to the beginning of the signpost so users can relate signposts to objectives. Avoid referring to complex concepts or terms when possible — both users and LLMs find them confusing.
  • Use proper HTML structure to create signposts so AI understands your content’s formatting. You can make the text look like a heading with some text decoration, but AI might not recognize it and struggle to find the right information. Use H1, H2, H3 tags and beyond so AI and assistive technology tools can read the information easily.
  • Implement the right navigation. Intuitive menu design and robust search functionality are vital to delivering a good customer experience. A customer should be able to quickly view category menus to find information they’re looking for and should be able to use the search functionality to find content with specific keywords. You should also organize information from the broadest to the most specific to make navigation easier.

Need some inspiration? Lyft’s knowledge base is a standout example. It all starts with a search bar — this is your easiest and most user-friendly option. Below, it displays popular articles. And, if you still can’t find what you’re looking for, it helps you narrow down the search by selecting a user-based category. It concludes with an extensive list of self-help topics.

An intuitive design and robust search are critical, but not enough. The most savvy companies use generative AI-powered customer service automation — an AI agent that can interpret a customer’s query and generate human-like responses.

Suppose your product is an FP&A SaaS. When a customer asks your AI agent “how to check the cash coming in and out of my business,” the AI agent understands the customer is asking for a cash flow statement. It uses the information in your knowledge base and uses natural language processing (NLP) to generate a response.

Step 4: Find knowledge base software

G2 has over 170 listings for knowledge base software. It’s hard to assess all solutions and select one, but we can give you a list of things to look for to make the choice easier. Find knowledge base software that’s easy to use, scalable, has an extensive feature set, and integrates with the apps in your tech stack. Software that's composable, like Contentful for example, are great options for scalability. Here’s a quick overview of features to look for in knowledge base software:

  • Integrations: Integrations help streamline workflows and save you and your customers considerable time. For example, integrating your AI knowledge base or knowledge base software with conversational AI customer service automation, like an AI agent , enables customers to get answers faster and improve the customer experience. Instead of digging out information manually, customers can ask questions to your AI agent. It will pull information from your knowledge base and generate a response for your customer.
  • Customizable UI: Knowledge base software with customizable UI helps create an easy-to-navigate interface. Your knowledge base software should allow you to customize CSS and HTML so you can optimize UX with a clean and intuitive design.
  • Collaboration and versioning: Authorized employees from various departments should have the ability to contribute information to the knowledge base. When you have multiple contributors, versioning is critical too. For example, when a team member updates a guide and makes errors, versioning will help you find those errors faster.
  • Analytics and reporting: AI-powered knowledge base software can collect a goldmine of data that can improve your customer service as well as your product. Built-in analytics and reporting features can convert collected data into insightful reports, so you can make data-backed decisions. But if your software doesn’t offer analytics and reporting, you can always integrate it with another tool that does, like an AI agent .

Step 5: Assign contributors

Developing your customer service team will be an important area of focus in the transition to an AI-first customer service organization . In an ideal world, you’d have dedicated employees managing your knowledge base — people with technical writing chops and strong collaboration skills.

In any case, you want to task someone who can write authoritative content and keep a pulse on company changes, product launches, and policy updates. While the knowledge base should have many stakeholders, there should be at least one person who can bring all this knowledge together and ensure it’s staying fresh and up-to-date. Maintaining this content means your AI is pulling from reliable sources, and in the end, it’s a mutually beneficial relationship — your AI will perform better, and you can use AI insights to continuously improve the knowledge base.

If you don’t have a dedicated person to oversee knowledge base management, select contributors from each department and make them responsible for managing a specific section of the knowledge base. Consider the following:

  • Domain expertise: Contributors need extensive knowledge of their respective domains to create authoritative content. You also need someone skilled in information architecture and technical writing to article information that comes from domain experts into snackable pieces of content that users can understand.
  • Communication skills: Contributors must be good communicators to deliver their ideas effectively. Look for contributors who can articulate information clearly and concisely.
  • Commitment: Contributors who are committed to building a comprehensive, high-quality knowledge base are more likely to help you maintain it. If a contributor doesn’t see enough value in building a knowledge base, they probably won’t be motivated enough to consistently work on improving the quality of information.

Step 6: Create and publish content

At this point, you can start producing articles, videos, and other resources. Be mindful of some best practices as you head into this next step of creating knowledge base content:

  • Prioritize critical information: For example, compliance information should appear before popular replies for customer service queries.
  • Clear, concise, and comprehensive: Steer clear of jargon. Use plain language and prioritize clarity. Aim for conciseness and comprehensive coverage — the aim is to cover all the necessary information using as few words as possible.
  • Add visuals: Big blocks of text are overwhelming. Break the monotony with images, GIFs, and diagrams. Add variety by creating video tutorials where it makes sense. For example, creating a walkthrough video is helpful when introducing a new feature. Note that generative AI and assistive technologies (like screen readers) only work with text, so you still want to minimize reliance on visuals and provide all of this information in text. Add alt text to images, captions to videos, and generate transcripts for audio.
  • Stay consistent: Establish editorial guidelines to maintain consistency in tone of voice and formatting across articles and videos. Create templates to help contributors stay consistent when creating content. For example, all feature update videos should start with an intro, mention the plans on which the feature is available, and then move on to the feature walkthrough.
  • Knowledge base SEO: Knowledge base SEO involves following SEO best practices: include the right keywords in the content, use descriptive headings, and maintain a consistent URL structure. Of course, there are a gazillion things on an SEO checklist so it’s best to seek professional help here.

Knowledge base best practices

There’s more to building a knowledge base than purchasing software and creating help content. Follow these best practices to ensure knowledge base success.

Deliver contextual and personalized solutions

Make no mistake, personalized interactions are not “nice to have.” Your customers want personalization. Research from McKinsey shows that 71% of customers expect companies to deliver personalized interactions and 76% are frustrated when that doesn’t happen.

You need to analyze data at scale to deliver highly contextual and personalized responses. That’s only possible with AI. The AI might look at the following data to deliver contextual and personalized responses:

  • Context: User actions indicate what a user needs help with. Was the user interacting with a specific feature for the past few minutes? If yes, they could be looking for help on that specific feature.
  • Behavioral data: A customer’s browsing history, usage pattern, and interactions with your app, website, or knowledge base.
  • User preferences: A customer’s physical location, preferences, and settings can help personalize responses. For example, if a user requests information on brokerage charges when executing trades via your investment platform, the AI can look at the default broker connected to the user’s account.

Integrate analytics

You can’t build a knowledge base and call it a day. Maintaining an extensive knowledge base — even an AI-powered knowledge base — requires consistent work.

Analytics identify popular topics and analyze search patterns, so when you know what users are searching for, you can double down on creating content that users need. Optimize your knowledge base content based on these insights to minimize exit rates and make the knowledge base more user-centric.

Update your knowledge base regularly

Whenever you introduce a new feature, update your product, change your branding, or undergo a noteworthy change in your company, share an update. Events like changes in branding or introduction of a new feature deserve fresh, dedicated articles. You should also monitor knowledge base search queries to find information users want but isn’t available in the knowledge base.

Aim to create one resource for each feature and topic. When you mention that feature or topic in another article, link it to the original resource. This way, you won’t have to update content on multiple pages when something changes.

Think about accessibility

Make it easy for all users, including users with diverse needs, to access the information in your knowledge base. Here are examples of things you can do to make your knowledge base accessible:

  • Add alt texts to all visuals
  • Add captions to videos for users with hearing impairment
  • Offer transcripts for audio content
  • Ensure enough contrast between text and background color
  • Set adequate font size to improve readability
  • Ensure customers who can’t use a mouse can navigate the knowledge base using just the keyboard
  • Add descriptive anchor texts for hyperlinks
  • Test the knowledge base with assistive technologies to identify potential issues and address them
  • Be mindful of other accessibility guidelines provided in WCAG 2.2 and Section 508 of The Rehabilitation Act, 1973 if you’re a U.S. entity

Share knowledge base content

Prospects often wonder how to use a feature instead of just hearing or reading about it. Satisfy their curiosity by showcasing knowledge base resources.

Link to knowledge base content when writing blog posts. Share articles that can solve someone’s problems on social media posts. And use knowledge base resources to illustrate concepts or showcase features during a webinar. For example, when you’re discussing a feature during a webinar, linking a video tutorial adds plenty of value for curious customers.

What tools should you include in your customer service knowledge base?

Knowledge base, AI agent, and forum are three key knowledge management tools. Using all knowledge management tools delivers the best customer experience. Let’s discuss how.

Knowledge base

Knowledge bases are a great self-serve tool. An information library where customers can browse for information they need without interacting with a support agent is a great way to match a modern customer’s need for self-serving. However, a knowledge base by itself leaves customers wanting more.

AirAsia, for example, has an extensive support section. The design is clean and intuitive and houses all the information flyers could need.

But to find information, customers would need to select a relevant category and maybe open a few resources. And when they can’t find an answer? Time to call support. That’s why AirAsia added an AI-powered customer service automation tool.

AI agents and scripted chatbot solutions

Scripted chatbots can understand customer queries and assist customers based on predefined workflows, answering FAQs and fielding queries for the customer service team based on knowledge base content.

AirAsia’s scripted chatbot helps customers find information faster and lowers support desk traffic. AirAsia passengers can use the chatbot to book flights, access real-time flight status, and update passenger information. After just a month of using Ada as their first line of support, AirAsia decreased the average wait time by 98% to less than a minute.

An AI agent takes it a step further, delivering mindful responses by reasoning through complex inquiries and collaborating with customers to provide the best resolution possible. An AI agent utilizes the information in the knowledge base and continuously improves.

Using natural language understanding (NLU) and machine learning, customer service automation tools can instantly interpret a customer’s query, scan your knowledge base for information the customer needs, and generate a response in natural language (thank you, generative AI). The customer doesn’t have to browse through piles of information or speak to a support agent — a win-win.


Forums allow customers to create threads on specific topics. If you’ve heard of sites like Quora and Discord, you’re probably familiar with the concept. A customer posts questions about your product in the forum and other customers and moderators post their responses to help the customer.

Creating a forum offers triple benefits: it reduces support desk traffic, helps build a community around your product, and is a great source of new content for the knowledge base.

Benefits of a knowledge base for customer service

By now, you probably know that a knowledge base is mission-critical for delivering a great customer experience. Let’s get more specific about what you gain by building a robust customer service knowledge base. Here are the key benefits.

Empowers customers to self-serve

Chances are you prefer checking in using the airline’s website or app. Your customers want to self-serve for basic tasks too, like recovering passwords or generating a statement. They want to access information without having to interact with a rep. The problem? Only a third of companies offer some form of self-service.

A knowledge base empowers your customers to find information independently. Empowering them to self-serve has perks for you too — your support traffic plummets and you spend less on handling customer calls. As a Harvard Business Review article explains:

“Corporate investment in self-service technologies has been enormously effective at removing low-complexity issues from the live service queue, and most companies we’ve studied report a steady reduction in such contacts over the past few years.”

Speaking of spending less time on calls…

Increases agent productivity

Knowledge bases can facilitate self-serve and reduce support desk traffic. But what about customers who can’t find the information they’re looking for? They call support. And when support agents spend valuable time browsing through resources, they can handle fewer calls per day. Even worse, the customer has to wait on hold, which can lead to a drop in CSAT scores.

An extensive customer service knowledge base with logically organized information and an AI agent enables agents to respond faster.

Picture this: a customer asks your support agent to help them integrate your SaaS with HubSpot. Since you’ve recently added the HubSpot integration, you need to look up the process. But instead of searching for information manually, you request the AI agent to find the information.

Reduces support costs

It’s easy to see how a knowledge base reduces support costs — fewer calls, fewer support tickets, and more productive support agents. In fact, Gartner estimates that conversational AI can help contact centers reduce agent labor costs by $80 billion by 2026. As Daniel O’Connell, VP analyst at Gartner, explains:

“Many organizations are challenged by agent staff shortages and the need to curtail labor expenses, which can represent up to 95% of contact center costs. Conversational AI makes agents more efficient and effective, while also improving the customer experience.”

Suppose you’re planning to hire five new agents ahead of peak season. You plan on paying each agent $60,000. You’re prepared to spend $300,000. Instead of taking on that operating leverage, you could just invest a few thousand dollars to create a knowledge base. It’s cheaper and you won’t have to fire anyone when demand drops.

Provides data-driven insights

Knowledge base software with built-in analytics collects data and generates reports summarizing key usage metrics. This data tells you what customers are searching for the most and helps identify gaps in your knowledge base content.

When you notice a high exit rate for a specific knowledge base page, look for gaps or confusing sections. Which section are they reading when they exit the page? Are you explaining the process with enough detail? Does the page have enough visual resources to support the text? Answer these questions to find the problem and fix it.

Future of customer service: The AI-powered knowledge base

Most of your competitors have a knowledge base. But AI can make yours stand out. Combining AI knowledge base software with an AI Agent gives you an edge, and your customers an experience they crave. While your competitors spend on hiring more support agents and training them, go full throttle to building an AI-powered knowledge base with Ada, where our mission is to transform customer experiences with AI.

The generative AI toolkit for customer service leaders.

Evolve your team, strategy, and tech stack for an AI-first future.

Get the toolkit