Create an FAQ model

Detect FAQ answers models use natural language processing (NLP) to understand customer inquiries and provide the correct answers in real-time.

Step one: Choose a topic for your FAQ model

Begin by choosing a single topic, such as “Shipping and delivery”, or “Returns” for your FAQ model.

By choosing a single topic, you can keep your model focused and organized. You can create more models for your other FAQ topics later.

FAQ answer models can only support 100 FAQ answer pairs. This means that each FAQ answer model can answer a maximum of 100 distinct questions.

Select at least two of your most frequently asked questions within the topic that you’ve chosen.

For example, if your topic is “Shipping and delivery”, your two most commonly received questions might be:

  • “How long does shipping usually take?”

  • “What countries do you ship to?”

Step three: Create question variations

Next, it’s time to generate variations of these FAQ questions. Think about the different ways in which your customers might ask the same question.

For example, the question “How long does shipping usually take?” might also be phrased like these questions:

  • “When will my order be here?”

  • “What are your delivery times?”

  • “When will my parcel get here?”

And the question “What countries do you ship to?” might be phrased like these questions:

  • “Do you offer international shipping?”

  • “Do you ship outside of the US?”

  • “Can I order from outside the US and still get it delivered?”

Come up with as many variations of each FAQ as you can think of. Ten is a good number to aim for!

Step four: Provide answers

For each FAQ question, make sure you have an answer that you want the automated system to provide.

For example, if your question is “How long does shipping usually take?”, your answer might be “Our standard shipping time is 2 working days. You can track your order by following the link in your order confirmation email.”

Step five: Define irrelevant content

Irrelevant content is any question that your customer might ask that doesn’t correspond to the FAQs or answers in your model.

By default, the model will automatically categorize a question as “irrelevant” if it can’t find a matching answer.

You can choose to add irrelevant content when you’re training your model, to provide it with examples of questions that you don’t want it to try and answer.

Step six: Test your model

Once you’ve built your model, you can test your model to see how well it works. If you don’t get the results you expect, you can go back to your FAQs and refine them to improve the accuracy of the answers the model provides.

Step seven: Use your generate FAQs answers

Now that you’ve created your generate FAQ answers model, you can deploy this model using the Agent Assistant FAQ Answers feature or the generate FAQ answers action for Flows.

Last updated