FAQPage Schema and Speakable Schema for Healthcare AIO
What Google actually does with FAQPage schema in 2026
Google reduced the visible FAQ rich-result expansion in 2023, and a lot of Indian SEO teams read that as "FAQ schema is dead." It isn't. Google narrowed the SERP display; the schema is still consumed by four downstream systems: AI Overview citation selection, People Also Ask expansion, voice-answer selection on Google Assistant and mobile, and internal query understanding. Removing FAQPage schema from a healthcare page after this narrowing produced measurable AIO citation losses on 40+ pages we A/B tested in Q1 2026.
The correct read of the 2023 change: Google took away the free advertising and kept the underlying signal. The teams that stopped shipping FAQ schema lost citation eligibility. The teams that kept shipping it — and tightened it — moved into the AIO citation slot as their competitors dropped out.
The three eligibility rules FAQ schema has to pass
- The question text is a real question. "About our services" is not a question. "How much does an IVF cycle cost in Delhi in 2026?" is. Google's AIO ranking model heavily weights question-form.
- The answer text on the page matches the schema payload word-for-word. Any material divergence causes Google to discard the schema under its structured-data policy.
- The answer is a self-contained paragraph of 40 to 90 words. Too short reads as thin; too long fragments the citation-eligible chunk.
Every FAQPage schema block on this website is generated from a validation layer that enforces these three rules at build time. When a page fails validation, it does not ship. That is the discipline that produces citation appearance rate — not the number of schema blocks.
The minimal FAQPage payload that works
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "How much does an IVF cycle cost in Delhi in 2026?",
"acceptedAnswer": {
"@type": "Answer",
"text": "An IVF cycle in Delhi in 2026 typically costs between 1.4 and 3.2 lakh, depending on the medication protocol, ICSI vs conventional IVF, and whether donor gametes are required. Standard first cycles at accredited centres cluster between 1.6 and 2.1 lakh, with add-ons like PGT-A and embryo freezing costed separately."
}
}
]
}
</script>
Notice the answer text is a single self-contained paragraph, includes qualifying numbers, and does not reference a previous section. That is what "citation-eligible chunk" means in practice.
What Speakable schema is and why it matters for healthcare
Speakable schema is a Schema.org extension originally scoped to news publishers, but the specification permits any content the publisher considers appropriate for text-to-speech playback. Google's voice-answer surfaces — Assistant, mobile voice search, and increasingly the voice mode of AIO on mobile — read Speakable-tagged passages preferentially.
For healthcare in India, this matters more than it did two years ago because a rising share of "cost of dental implant near me" and "what are the symptoms of PCOS" queries are being spoken aloud on mobile phones in Tier-2 and Tier-3 cities. Text-to-speech answers those queries by reading a single passage. Tagging the passage explicitly with Speakable schema tells Google, "this is the passage designed to be read aloud."
The Speakable pattern that pairs with FAQPage
The compounding move is to tag the same 40 to 90-word answer paragraphs — the ones already inside your FAQPage block — with Speakable. The two schemas do not conflict; they mark the same passage for two different downstream consumers.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "WebPage",
"speakable": {
"@type": "SpeakableSpecification",
"cssSelector": [".icg-faq-answer"]
}
}
</script>
Every FAQ answer paragraph on your page carries the .icg-faq-answer class. Speakable picks them up. FAQPage carries them as machine-readable Q&A. Both signals point at the same passages, which is exactly what Google's model wants to see.
How many FAQ questions is right for a healthcare page
Between three and eight. Below three, Google reads the FAQPage block as thin. Above eight, the questions start cannibalising each other — two nearly-identical questions each with a 60-word answer will produce zero AIO citations, because Google cannot decide which one is canonical. The correct discipline is to pick the five to seven most distinct questions from your Search Console, PAA, and call centre logs, and write one clean answer to each.
The healthcare-specific compliance layer
Healthcare content in India carries constraints most SEO teams do not internalise. The IMC Ethics Regulations 2002 and the NMC 2026 advertising guidelines prohibit guaranteed-outcome language, testimonials, and comparative claims. Answers inside FAQPage schema that violate these rules are not just SEO problems — they are compliance exposure that persists in Google's index long after you have quietly deleted the on-page copy.
The safe-answer template we run: state the general medical context, provide a factual range where numbers apply, cite the source or professional body where possible, and close with a "consult a qualified specialist" line. That template passes AIO citation eligibility and NMC review simultaneously. Anything more claim-heavy risks both.
The measurement layer — what changes and when
- Week 1 to 3 — Rich Results Test passes for FAQPage and Speakable. GSC "Structured Data" report shows the new items. No visible SERP change yet.
- Week 4 to 8 — People Also Ask appearances rise on the exact questions used in the schema. This is the leading indicator of AIO citation eligibility.
- Week 8 to 14 — First AIO citations appear on the target queries. Manual logging is the only reliable measurement here; GSC does not distinguish AIO citations from blue-link impressions.
- Week 12 to 20 — Branded query lift becomes visible in GSC as AIO citations bring the brand to the reader's awareness earlier in the funnel.
What kills FAQPage schema effectiveness on healthcare pages
- Duplicating the same FAQ block across dozens of pages. Google collapses duplicate schema and picks one canonical page. The rest become schema noise.
- Copy-pasted answers from competitor sites. Instantly detected and discarded, and puts you at a topical-authority disadvantage on the query.
- Nested marketing links inside answer paragraphs. The AIO extractor prefers clean prose. Answer paragraphs with 3 internal links do not get pulled cleanly.
- Answers that reference "as shown in the video above" or "as we discussed in the previous section". Not self-contained, not citation-eligible.
The pairing with entity SEO and video
FAQPage and Speakable schema are the last-mile signal on a page that has already earned entity resolution and topical depth. Ship them on a page from an unresolved brand and they are noise; ship them on a page from a resolved entity with corpus depth and they compound. This is why our AIO programme sequences entity work first, corpus depth second, and schema third — not the reverse. And this is why YODA-produced YouTube on the same topics as your text pages accelerates AIO citation appearance: video citations and text citations reinforce each other inside the same AIO block.
Get your healthcare pages schema-audited
We will map every AIO-eligible page on your site, validate the FAQPage and Speakable payloads, flag every on-page/schema mismatch, and lay out the fix sequence. Fixes typically produce citation appearance inside 60 days.
Book a free audit WhatsApp ICG