Creating FAQ from reviews
Use AI to generate frequently asked questions from customer reviews
Use customer reviews to automatically generate frequently asked questions (FAQ) content that addresses real customer questions while boosting SEO through long-tail keywords and rich snippet eligibility.
Don't have customer reviews?
No worries! By leveraging Emfas's Deep Research feature you can find and generate FAQ on existing or similar products. Read more about it here.
Step 1: Importing Reviews
-
Create a plain text Reviews attribute to store your customer reviews. See creating attributes for details.
-
Upload your customer reviews as a CSV with two columns: the product identifier (handle in Shopify, product number in Centra) and the reviews for that product as seen in the following example:
| Product Handle/Number | Reviews |
|---|---|
| burton-yeasayer | So playful and fun! Perfect for messing around in the park. Lightweight and easy to spin. My 360s have never been cleaner. |
| burton-carver | This thing RIPS on hardpack. Best edge hold I've ever experienced. Stiff and responsive, not for beginners, but if you know how to drive a board it rewards you big time! |
For detailed import instructions, see CSV import. The more reviews you include the better, you can also include your own responses or answers if available.
Step 2: Setting up the FAQ attribute
- Create a plain text or HTML FAQ attribute
- Enable Generatable on the attribute
- Add instructions to the attribute, for example:
Generate an FAQ section based on customer reviews. Analyse the 'Reviews' attribute and identify 6-8 common questions or concerns customers have. Write clear, helpful answers that address each one. Avoid mentioning things like "customers say...". Format using HTML: wrap questions in bold
<b>, answers in plain text, with one newline<br>after the question and two newlines<br><br>between each Q&A pair.
Don't forget to save your changes!
Step 3: Generate FAQ using AI
-
Use filtering to find all products with reviews
-
Create a proposal for the 'FAQ' attribute
-
Review and publish the FAQ
For more info on generating with AI you can check out this guide.
Best practices
-
Include diverse reviews: Mix positive, negative, and neutral reviews. Complaints often reveal the most useful FAQ topics (sizing, durability, compatibility).
-
More reviews = better FAQ: The AI needs enough data to identify recurring themes.
-
Include your own responses: If you've already answered customer questions in review replies, include those. The AI will use them to craft accurate answers.
-
Review the output: AI-generated answers should be fact-checked against your actual product specs, return policy, and support documentation.
-
Update periodically: As new reviews come in, regenerate FAQ to capture emerging questions or concerns.
-
Consider language: If you have reviews in multiple languages, you may want to generate FAQ per language or consolidate to your primary language first.
Generating FAQ using Deep Research
Don't have customer reviews? Use Deep Research to find FAQ content from across the web.
Great for multi-brand retailers
If you sell products from established brands, Deep Research can find existing customer questions from the manufacturer's site, Amazon, Reddit, and other sources; giving you rich FAQ content without needing your own review data.
- Create a plain text or HTML FAQ attribute
- Enable Generatable on the attribute
- Enable Deep Research on the attribute
- Add instructions, for example:
Research common questions and concerns customers have about this product. Search for existing FAQ, reviews, and Q&A from retail sites, forums, and the manufacturer. Compile 6-8 relevant questions with accurate, helpful answers. Format using HTML: wrap questions in
<b>tags, answers in plain text, with two<br>tags between each Q&A pair.
- Create a proposal and generate FAQ.