TL;DR
A product page without an FAQ sends shoppers to your support inbox or to a competitor. AI can draft a solid first version in about ten minutes if you feed it your real policies. Three things changed by June 2026 and they reshape the goal:
- Google removed FAQ rich results from Search on May 7, 2026. The schema still validates, but it no longer earns the expandable snippet on product pages, so write the FAQ for humans and AI answer engines, not for a star in the SERP.
- The global average ecommerce conversion rate sits around 2.7% (2.5-3.0% depending on industry), and most of that friction lives on the product page, not at checkout. An FAQ that kills a specific objection is cheap conversion work.
- Shopping assistants in ChatGPT and Gemini now read your page text to answer “does this fit a US 10?” A clear, factual FAQ is what they quote.
Why a product FAQ still earns its space
For years the pitch for an FAQ section was the rich result: that expandable block of questions in Google Search. That pitch is gone for commercial pages. Here is the exact timeline so you can plan around it.
| FAQ rich result milestone | Date | What it means for you |
|---|---|---|
| Rich results stop appearing in Search | May 7, 2026 | Product-page FAQs no longer get the SERP snippet |
| FAQ report + Rich Results Test support dropped | June 2026 | You can no longer preview the feature in Google’s tools |
| Search Console API support removed | August 2026 | Scripts pulling FAQ data from the API will return nothing |
The FAQPage schema type itself is not deprecated. Leaving it on the page throws no error, triggers no manual action, and does not hurt rankings, and Bing plus most AI crawlers still parse it. So keep the markup if you have it, but stop treating the FAQ as an SEO trophy. The payoff now is on-page: a shopper who finds the answer to “will this ship to Canada before the holidays?” in two seconds is far more likely to buy than one who closes the tab to email you.
When AI is the right tool for this
- You already have a product description, specs, and a written shipping and returns policy to paste in.
- You sell at least 10 similar SKUs and want one consistent FAQ structure across all of them.
- You know your top three to five buyer hesitations from real emails, chat logs, or reviews.
In that situation AI compresses an hour of writing into roughly ten minutes per page. If you do not yet have the raw policy text, fix that first. AI cannot invent your return window for you, and you do not want it to.
The one rule: never let AI guess a policy
Ask a model “what is your return window?” without telling it, and it will confidently write “30 days.” That guess lands on a live product page and stays there until a customer holds you to it. Always paste your real policy text into the prompt, and always read every generated answer against your terms page before publishing. Treat the AI as a fast first-draft writer, not a policy author.
What to feed the model
- Product name, category, and a 3-5 line description
- The top three to five questions from real customer emails or chat transcripts
- Your shipping, returns, warranty, and care policy as plain text
- One competitor FAQ you admire, pasted in as a tone reference
Which AI tool to draft with
Any current frontier model handles this well. The differences that matter for FAQ drafting are how much policy text you can paste in and whether the tool can read a live competitor page.
| Tool | Tier to use | Why it fits FAQ work (as of June 2026) |
|---|---|---|
| ChatGPT (GPT-5.5) | Plus, $20/mo | Strong default writer; in-app context ~320 pages is plenty for one product’s policies |
| Claude (Sonnet 4.6) | Pro, $20/mo | 1M-token context, clean structured output, careful about not inventing facts |
| Gemini 3.1 Pro | Google AI Pro, $19.99/mo | 1M context plus can pull a live competitor URL for tone reference |
For a single product page, the free tier of any of these is enough. Pick the paid tier only when you are batching dozens of SKUs in one session and hitting message limits. See our guide to product FAQ prompts for ready-made variants of the template below.
Copy-ready prompt
Paste this and replace each bracketed placeholder with your real text.
You are writing the FAQ section for one product page.
Product: [product name]
Category: [category]
Description: [3-5 line description]
Shipping policy: [paste your real shipping text]
Return policy: [paste your real returns text]
Warranty / care: [paste your real warranty and care text]
Top buyer hesitations from real emails: [list 3-5]
Write 8 FAQ entries. For each entry:
- Question: one line, phrased the way a buyer actually types it (lowercase and casual is fine).
- Answer: 2-3 sentences. Lead with the direct answer, then one supporting detail.
Cover, in this order: sizing and fit, materials, shipping time and cost, returns,
care, warranty, comparison to the obvious alternative, and the single biggest
hesitation from the list above.
Hard rule: do not invent any policy. If a policy detail is missing from what I
gave you, output the literal text [POLICY MISSING] in that answer instead of guessing.The [POLICY MISSING] flag is the safety valve. Search the output for it before you publish, and fill each one from your real terms.
A worked example
Say you sell a merino wool base layer and your top hesitation from emails is itchiness. A weak AI answer reads: “Our base layer is comfortable and made from high-quality materials.” That is generic filler a buyer ignores. A strong answer, grounded in your specs, reads: “It is 100% 17.5-micron merino, which is fine enough that most wearers feel no prickle against skin. If you have reacted to coarser wool before, the 17.5-micron grade is the same fineness used in next-to-skin sportswear.” The difference is one number pulled from your product spec, and it is the line that converts.
How to check the output before publishing
- Read every answer out loud. If you stumble, the sentence is too long. Cut it.
- Cross-check each policy claim against your real terms page, not your memory.
- Search the
[POLICY MISSING]flag and resolve every hit. - Open your last 30 days of support inbox and add any recurring question the FAQ missed.
Common mistakes
- Generic answers that would fit any product in the category. Specifics convert; platitudes do not.
- Answering the questions you wish buyers asked instead of the ones they actually send.
- Placing the FAQ above the buy button. It should support the decision, not stand between the shopper and the cart.
- Letting it go stale. Policies and SKUs change; review the FAQ each quarter.
- Stuffing the FAQ purely to trigger schema. That ship sailed on May 7, 2026. Write for the reader.
Keep it fresh
Book a recurring 20-minute block each quarter to fold the last quarter’s support tickets into each FAQ. Watch click depth on the FAQ section in your analytics: a question that gets clicked a lot probably belongs higher on the page, or in the product description itself. If you also run a help center, the same source material feeds it. See help center FAQ with AI for that workflow.
FAQ
- Do FAQ rich results still work in Google? No. Google stopped showing FAQ rich results in Search on May 7, 2026, and is removing tool and API support through August 2026. The FAQPage schema still validates and is still read by Bing and AI crawlers, so keep your markup but do not expect a SERP snippet from it.
- How many FAQ entries should a product page have? Six to ten is the sweet spot. Fewer than five looks thin; more than twelve buries the answers that actually move a buyer.
- Should FAQ answers link to policy pages? Yes. Keep each answer self-contained, then link to the full policy for shoppers who want the fine print.
- Can I reuse the same FAQ across similar SKUs? Share the structure, but rewrite the sizing, materials, and comparison answers per SKU. Those are the lines a buyer reads most closely.
- Will AI shopping assistants use my FAQ? Yes. ChatGPT and Gemini shopping flows read on-page text to answer buyer questions, and a clear, factual FAQ is exactly what they quote, so accuracy matters more than ever.
Related
- Prompt library: FAQ section prompts
- Prompt library: product FAQ prompts
- Help center FAQ with AI
- Fix invalid FAQ schema
Tags: #E-commerce #Workflow