# How to create a good FAQ for Lyro?

Section: Lyro - AI Agent

Description: Write clear FAQ answers and use Lyro conversations to find missing knowledge.

Verified against Tidio panel 12.331.0

Last updated: 2026-11-10

Source: https://help.tidio.com/articles/good-faq-for-lyro/

Frequently Asked Questions are an effective way to provide helpful information to users. Most businesses include an FAQ on their website for visitors to see, but you can also create automated FAQs in Tidio. In this article, let's focus on the conversational AI agent called **Lyro** - it is an automated assistant whose knowledge is based on your FAQ.



By structuring the information in a question-and-answer format, Lyro can understand user queries and respond with relevant answers; a well-structured FAQ is key in this case. We will go over some tips on creating effective FAQs for your Lyro AI agent.







**In this article, you'll learn:**



- [The main ideas behind a good FAQ](<https://help.tidio.com/articles/good-faq-for-lyro/#h_01HBFT0RR75E35TEGMQCB2GQAC>)
- [How to handle edge cases](<https://help.tidio.com/articles/good-faq-for-lyro/#h_01HBFT0RR7X7KC5K0HC2R50AMK>)
- [How to include product-related information](<https://help.tidio.com/articles/good-faq-for-lyro/#h_01HBFT0RR7342TQ8Q092PZGDSP>)
- [What to avoid when creating your FAQ](<https://help.tidio.com/articles/good-faq-for-lyro/#h_01HBFT0RR753240T640GY95Y2W>)
- [What some examples of good (and bad\!) FAQs look like](<https://help.tidio.com/articles/good-faq-for-lyro/#h_01HBFT0RR7W9BD9R8V615G2J0D>)
- [How to improve Lyro's performance](<https://help.tidio.com/articles/good-faq-for-lyro/#h_01HC0N65E94GFGA77QFQ1J5QAC>)



[Video](<https://help.tidio.com/article-media/good-faq-for-lyro/walkthrough.mp4>)



Silent walkthrough. Open the source review drawer on a demo Lyro email reply and review its Q&A and website sources.







If you'd like to learn more about what Lyro is and how it works in general, please check out [this article](<https://help.tidio.com/articles/lyro-main/>)\!











## Main ideas







### Focus on User Intent



A good FAQ targets your users' questions and intents while they're interacting with your AI agent. Analyze customer inquiries, support tickets, forum discussions etc. to identify common requests, questions, and pain points. Write your FAQs to directly address these user intents.







### Use Natural Language



Structure questions conversationally, as users would naturally ask them - don't use robotic or overly formal language. This helps the Lyro's language model understand and respond properly to users.  
For example, use "How do I reset my account password?" and not: "Password reset procedure explanation request."







### Keep Questions Short



Long or overly detailed questions can be confusing for users and challenging for the AI to comprehend. Keep your questions simple, clear, and to the point.







### Maintain a Consistent Format



Keep questions grammatically consistent and answers in a uniform format. This helps Lyro recognize patterns more efficiently.







### Test Extensively



Thoroughly test your FAQs with real user queries. Identify gaps or problematic areas and keep improving. Make sure to expand your FAQs as more usage data comes in.







### Update Regularly



View your FAQs as an evolving, living document. Add new questions and answers whenever new user intents arise; remove outdated content; keep it current and relevant. You can use the **Search question** field to look for specific words in your current data sources.



💡



**Please note:** FAQ size alone does not tell you how many conversations Lyro will resolve. Start with the questions your customers ask, test the answers, and review conversations to find gaps.











## Handling edge cases



It is a good idea to include less common questions in your FAQ as well. You may find that your users will rather often come with slightly unusual problems; you can keep Lyro prepared for such situations. Here are some tips for handling edge cases:







### Anticipate edge cases when analyzing user conversations



Look for unusual questions, niche issues, or special scenarios that come up. Document these as potential FAQ topics.







### Use a consistent format



Even for obscure issues, follow the same clear Q&A structure used for other FAQs.







### Provide context before details



Briefly explain what scenario the edge case question refers to before diving into specifics. This helps Lyro understand.







### Focus details on resolution



In the answer, get to the solution steps first, before explaining background. Users in edge situations are looking for help.







### Point to other resources



If an edge case requires lengthy troubleshooting, provide links to knowledge base articles or other materials in the answer.







### Monitor conversations



As new edge cases emerge, continuously add to this section of the FAQ.



The goal is to anticipate and document edge cases without letting them dominate the FAQ. Maintain a user-focused and problem-solving approach for edge case content, and point users to human support when automation can't help anymore.











## How to include product questions in the FAQ?







### Keep it high-level



Avoid too much technical detail or spec sheets. Focus on user goals and key product capabilities.







### Focus on most popular features



FAQs should cover the product features customers ask about most frequently. Be selective about going in-depth.







### Use common language



Explain features using vocabulary a typical user would understand, not complex industry jargon.







### Link to product docs



For users who want more details, include links to user manuals, specs sheets, tutorials, etc. in the FAQ answers.







### Keep updated



If the product changes over time, update the relevant FAQs to reflect the latest features and functions.







### Indicate limitations



Be clear about which product questions the agent can and cannot address in its current state.







The goal is to provide users with a quick overview of the product's main functions and capabilities, not overwhelm with information. Detailed articles and tutorials are better suited for self-service browsing, e.g in your help center. Your FAQ should make the product feel approachable.











## What to avoid?







### Duplicate questions



If the same intent is already covered, don't repeat it.







### Too much detail



Avoid long, complex answers when a shorter explanation will do. Save the details for separate docs.







### Marketing fluff



Don't just praise your product's features; focus on addressing real user needs.







### Offensive material



FAQs should follow the same etiquette guidelines and content policies as other user-facing materials.







### Legal information



Lengthy terms & conditions, privacy policies, etc. are better placed on separate legal pages.







### Anything requiring ongoing updates



Avoid FAQs about fluid things like pricing, availability, etc. Link to pages you can update dynamically instead.







### Anything promoting unethical behavior



Don't provide guidance on how to misuse services, hack accounts, spread misinformation, etc.







### Questions the agent truly can't handle



Be honest about Lyro's limitations. Don't try to fake it.







Your FAQ should give customers helpful, accessible information. Avoid content that degrades the experience or damages your users' trust. Keep it focused, friendly, and genuinely useful.











## What could you forget?







There are aspects worth double-checking while you're creating and enhancing your FAQ. Below you will find a helpful list of things you may have missed in the process:







- Forgetting to analyze real user conversations - Neglecting to review support tickets, forum posts, chat logs, etc. means missing important user intents.
- Using inconsistent formatting - Randomly formatted FAQs confuse Lyro's ability to recognize patterns.
- Not testing thoroughly - Test the FAQ in the Playground before launch to identify gaps.
- Failing to plan updates - FAQs get outdated quickly. Not planning your reviews and updates will cause you to miss opportunities.
- Assuming one FAQ set will suffice - Different user segments may need tailored versions of the FAQ focused on their needs.
- Forgetting internal resources - Customer support and product teams can suggest additional helpful pre-emptive FAQs. Consult with your co-workers\!







Putting thought into the full user journey can help identify areas that are easy to miss when initially compiling an FAQ. Considering future maintenance and expansion also leads to more robust FAQ documentation.











## Examples







Let's take a look at some examples of good and bad question-and-answer pairs for Lyro's FAQ.







## Good Q&As:



Q: How do I reset my password?  
A: You can reset your password by going to your account settings and clicking "Reset Password." Follow the prompts to create a new password.



Q: What payment methods do you accept?  
A: We accept Visa, Mastercard, American Express, Discover, PayPal, and Google Pay. You can update your payment information in your account settings.



Q: My package hasn't arrived yet. What should I do?  
A: If your package has not arrived within 7 business days, please contact our customer support team at [support@company.com](<mailto:support@company.com>). We will investigate the situation and provide an update within 48 hours.







## Bad Q&As:



Q: Reset account credential data  
A: Go to settings, reset password



Q: Payment type options?  
A: V M A D P G



Q: Parcel status explanation request. Estimated time of arrival exceeds expected window.  
A: Contact assistance team. Wait 2 days for conclusion.



The bad examples use unnatural language, lack context, and have formatting issues. The good examples use conversational language, include context, and are structured in a clear Q&A format.











## Improving Lyro's performance







While you may already have all the right tools at your disposal, having a clear step-by-step process can be really useful for improving how Lyro handles chats. Feel free to use the steps below as a guide, and optimize your bot's performance over time\!







1. Add your first FAQ - Do not worry if Lyro won't be able to answer all of your users' questions fully right away. This is the very first step in Lyro's learning process; with the first data source you add, there will definitely be room for improvement\!
2. Check Lyro's conversations - Whenever Lyro transfers a chat to your agents for human assistance, that is a great opportunity to check how Lyro handled the user's questions so far. Take note of what Lyro did incorrectly, and consider how you can update the FAQ. Try to make Lyro understand the user's question or issue more accurately.
3. Update Lyro's data source - Once you've identified where Lyro's knowledge was lacking, add a new, appropriate question-and-answer pair to the FAQ (or add more detail to an existing pair). Please note: if Lyro fails to answer a specific question, you can use the **Create answer** button precisely where it happened, directly inside that conversation - and update Lyro's knowledge without accessing Lyro's configuration\! You should also see a **Review sources** button whenever Lyro provides a reply according to its current Q&A. Once clicked, you will see the sources used on the right side of the panel.
4. Test Lyro's knowledge - Now that you've updated the data source, use the Playground to check if Lyro performs better. Try asking a problematic question in different ways and see if Lyro is able to handle the conversation on its own.
5. Repeat steps 2-4 regularly - Make sure to check how Lyro performs whenever you get the chance; update the FAQ as needed, and do your own testing once more. This process is sure to help improve how Lyro assists your users\!











Open a Lyro conversation and select **Review sources** beside a reply to see the knowledge it used:

![A demo Lyro conversation with the Review sources drawer open, showing the Q&A used for the selected reply.](<https://help.tidio.com/article-media/good-faq-for-lyro/13.1.png>)

## Further reading



Use [Lyro Suggestions](<https://help.tidio.com/articles/lyro-suggestions/>) to review missing knowledge, or learn [how to import knowledge into Lyro](<https://help.tidio.com/articles/import-knowledge-into-lyro/>).
