Here's What Actually Changed (And What Matters for Your website)
FAQ Schema Isn't Dead — Just the Rich Result. Here's What Actually Changed (And What Matters for Your App)
May 2026
On May 7, 2026, Google removed FAQ rich results from search. The FAQPage structured data type itself? Still valid. Those are two different things—and the difference determines whether you need to rip out your schema or leave it alone.
If you work on search visibility, content indexing, or AI retrieval pipelines, this distinction matters. This is the practitioner's breakdown: what changed, what didn't, why it matters, and what to do based on what you actually build.
The Critical Distinction: SERP Feature vs. Schema Type
Let me be direct: the SERP feature died. The schema type did not.
What Died on May 7
FAQ rich results—those expandable Q&A dropdowns that lived below search listings—no longer appear in Google Search. For non-authority sites, this feature was functionally dead since August 2023 anyway. The May 2026 announcement was closing a door that was already mostly shut.
Timeline for Google Search Console removal:
- FAQ rich result feature: Removed May 7, 2026 (effective immediately)
- Search Console FAQ reporting: Removed June 2026
- Rich Results Test for FAQ: Removed June 2026
- Search Console API FAQ endpoint: Removed August 2026
What Stayed
FAQPage structured data is still valid markup. Google's documentation states explicitly: you don't need to remove it. Structured data that isn't being rendered does not cause indexing problems.
At the crawler layer, multiple systems still consume this markup:
- Bing (powers ChatGPT search and Microsoft Copilot grounding)
- PerplexityBot
- Voice assistant indexers
- RAG crawlers for generative systems
None of those systems have published how they weight FAQPage markup in their retrieval or generation pipelines. But they're all parsing it.
Why This Happened: The Pattern to Understand
This isn't the first time Google has done this.
May 2019: FAQ rich results launch. Immediate, easy SEO win.
August 2023: Google restricts FAQ rich results to government and health sites only. For everyone else? Feature functionally disappears.
May 2026: Google removes the feature entirely, even for eligible sites.
The pattern: Open a feature → observe industry-wide scaling → restrict eligibility → remove feature entirely → schema type remains valid.
Same sequence happened with HowTo rich results (deprecated September 2023 after mobile-only restriction).
Why Now? The Theories
Several credible explanations exist:
Spam saturation – Templated FAQ blocks at scale signal low-quality content. Google saw 168,000+ articles with templated FAQ schema and responded.
LLMs don't need the schema layer – Modern language models parse natural prose just fine. A clean Q&A under a
<h2>heading with proper structure extracts as easily from plain HTML as from JSON-LD.AI Mode strategy shift – Google is consolidating SERP features into AI Mode follow-ups. Removable rich results (HowTo, FAQ, news carousels) migrate toward AI Overview handling instead.
Quality signal correlation – FAQ schema abuse correlated with Helpful Content Update impacts. Removing the feature pushes away the gaming.
The action is the same regardless of cause: Stop using FAQ schema as a shortcut. Start treating it as an honest description of pages that genuinely contain question-and-answer content.
The Architecture: Where FAQ Schema Sits in the Visibility Stack
Stop thinking of FAQ schema as a single layer. It exists on multiple floors of the visibility model:
Floor 1: Entity Foundation
NAP consistency, Bing indexability, Wikidata, Technical SEO
- Your entity can be found and correctly identified
- Nothing above this works without it
Floor 2: Content Extractability
Structured data, Schema markup, Machine-readable answers
- Crawlers (Googlebot, Bingbot, RAG pipelines) can parse, extract, and quote your content
- FAQ schema lives here. It's one tool among several for making content machine-readable.
Floor 3: Trust & Editorial Selection
Named sources, third-party citations, editorial corroboration
- AI systems have enough trust to recommend you, not just retrieve you
- Studies show 92% of Google AI Overview citations come from earned media, not owned content
Floor 4: Agentic Execution
MCP, WebMCP, Callable Tools, Governance
- Future layer: AI agents can transact on a user's behalf
- Being built now; most teams aren't here yet
The FAQ rich result was a bridge between Floor 2 and Floor 3. That bridge is gone. The floors themselves haven't moved.
What Changed on Your Site (If Anything)
Direct Impact by Use Case
| Type | Impact | What You Should Do |
|---|---|---|
| Regulated enterprise (gov, NHS, large health) | Real but small | Export GSC data before June; preserve schema where accurate |
| B2B SaaS | Minimal loss | Audit FAQ blocks; remove templated ones; keep schema on genuine Q&A |
| Professional services (law, finance) | Negligible loss | Convert templated FAQ blocks to editorial content; keep schema where appropriate |
| E-commerce / retail | Some loss | Stop adding FAQ schema for CTR; rebuild around Product schema + UGC |
| Local SME | Symbolic only | Skip the panic; focus on Floor 1 (Bing Places, NAP consistency) |
The Real Story
If you had FAQ rich results showing in Google Search in April 2026, you've already lost that traffic by May 7.
If you didn't have them showing (most non-health, non-government sites), nothing changed for you.
The decision question isn't "should I remove all FAQ schema?"—it's "should I remove the schema from this specific page?"
Decision Framework: Keep or Remove?
Ask three questions for each page with FAQ schema:
┌─ Is the FAQ visible on the page? ──► NO ──► REMOVE
│ YES
│ ▼
├─ Are the Q&As genuinely useful? ──► NO ──► REWRITE
│ (real questions, real answers, YES
│ not templated filler?) ▼
│
└─ Is this templated across ──► YES ──► AUDIT
many pages with similar text? (keep on pages where it fits,
remove from filler pages)
NO
▼
KEEP
Question 1: Is It Actually Visible?
If the Q&A only exists in your JSON-LD and users don't see it on the rendered page, it was added for the rich result. The rich result is gone. Delete it.
Google's guidelines always required FAQ content to be visible to humans. Pages that quietly violated that were never truly eligible.
Question 2: Are the Q&As Genuine?
Real user questions with substantive answers. Not keyword research output. Not marketing copy disguised as answers.
If the FAQ reads like templated boilerplate, fix the content. Schema follows content quality. Keeping mediocre schema doesn't fix a mediocre page.
Question 3: Templated Across Your Site?
If the same FAQ block appears on 15 service pages with minimal variation, that's spam-adjacent. Audit aggressively.
Keep it on pages where the FAQ genuinely describes that page's topic. Remove it from pages where it's footer boilerplate.
Implementation Timeline
This Week (1-2 hours)
- Export your FAQ rich result performance data from Search Console before June removal
- Run a before/after comparison on pages that had FAQ dropdowns showing
- Crawl your sitemap to identify all pages with FAQPage schema
This Month (half-day)
- Run the three-question test on each FAQ page
- Categorize into: Keep / Audit / Rewrite / Remove
- Implement removals immediately
- Create tickets for rewrites (these are content improvements, not schema-stripping)
This Quarter
This is where it gets strategic. The effort that went into "FAQ schema for visibility" should move upstairs in the building:
- Floor 1: Bing Places listing, Wikidata entry, Google Business Profile consistency, NAP alignment
- Floor 2: Honest structured data that accurately describes your page (CITATE framework: declarative openings, defined terms, named statistics with sources)
- Floor 3: Become a named source in editorial coverage of your domain
The research is clear: 82-92% of AI citations come from earned media. That's a Floor 3 problem. Most teams are still working on Floor 2.
Technical Debt
If your monitoring dashboards pull FAQ rich result metrics from the Search Console API, you have until August 2026 to redesign them. After that, those metrics don't exist.
Replace feature-specific tracking with:
- Total page CTR (not feature-segmented)
- AI citation tracking across platforms
- Entity mention tracking (Wikidata, knowledge graph)
The Bigger Pattern: Every Visibility Shortcut Has a Half-Life
This episode teaches something larger: extraction shortcuts get withdrawn.
- Sitelinks: Launched, scaled, restricted, deprecated
- Breadcrumb rich results: Same arc
- HowTo rich results: Same arc
- FAQ rich results: Same arc
The pattern isn't about whether structured data is valuable. It's about whether any specific shortcut, once scaled as a shortcut, survives.
What Survives
The work that compounds:
- Entity foundation (consistent, corroborated identity)
- Honest content extractability (your page actually contains what schema describes)
- Editorial citation (earned coverage)
- Substantive on-page evidence (named statistics, attributable claims, sources)
These don't get withdrawn because they're not features. They're properties of the page itself. AI systems either find a corroborated claim or they don't. Whether Google renders that as a rich result, AI Overview, follow-up suggestion, or plain text doesn't change what's actually there.
FAQ schema as an honest description of a page with genuine Q&A? Durable.
FAQ schema templated across 20 service pages because you read a "FAQ schema for GEO" guide? Removed in May 2026.
For AI/LLM Developers Specifically
If you're building retrieval-augmented generation systems, AI search, or generative AI products:
Schema markup is Floor 2 work. Necessary but not sufficient for ranking or recommendation.
The real signal lives on Floor 3: editorial citations and corroborated identity. That's where AI recommendation eligibility comes from.
Honest Floor 2 markup still matters because it improves extraction accuracy. But extraction accuracy isn't the bottleneck for most sites. Editorial trust is.
Parse what's there. If a page has FAQPage markup and the content matches, extract it. If the schema is decorative, treat it as noise. Most RAG systems do this already via confidence thresholding.
What Not to Do
❌ Don't strip all FAQ schema from your estate. That's treating a content decision as a markup task.
❌ Don't assume "schema is dead." The feature died. The markup type didn't.
❌ Don't overindex on the SERP impact. For most non-regulated sites, you had no SERP impact since 2023.
❌ Don't migrate effort to other rich results. HowTo, Breadcrumb, Product schema have their own half-lives coming.
✅ Do: Audit which FAQ blocks are genuine and which are templated.
✅ Do: Keep schema on pages where Q&A is the actual page format.
✅ Do: Move effort from "schema for visibility" to "earned media for authority."
The Take-Home for Builders
The lift (AI Overviews, AI Mode, Perplexity, ChatGPT search grounding) is being built. The temporary ladder (FAQ rich results) has been removed. The building hasn't moved.
Floor 1: Entity foundation. Still matters.
Floor 2: Content extractability. Still matters.
Floor 3: Editorial trust. More important than ever.
Floor 4: Agentic execution. Coming soon.
FAQPage schema is Floor 2 infrastructure. Use it where it's honest. Remove it where it isn't. Move the bigger effort upstairs.
Top comments (0)