AYUSH ministry guidelines — ayurveda and wellness ChatGPT Ads compliance
Ayurveda, yoga, unani, siddha and homeopathy brands carry a compliance layer no other healthcare vertical does: a named disease list that cannot be paired with a cure claim, drug licence linkage for every formulation mentioned, and state Ayush directorate rules that vary by geography. Inside a ChatGPT Ad, where the assistant can paraphrase a claim mid-conversation, that layer needs to be built into the source claim, not bolted onto the submitted copy.
Three instruments, one restriction at the core
The foundational document is the Drugs and Magic Remedies (Objectionable Advertisements) Act, 1954, which lists specific diseases and conditions — including diabetes, cancer, sexual disorders, and a defined set of others in its schedule — for which no advertisement may claim a cure, mitigation, treatment, or prevention, regardless of the system of medicine making the claim or the medium carrying it. The Act was written decades before conversational AI existed, but its language attaches to the act of advertising a cure for a scheduled condition, not to any particular format, which is why it reads on a ChatGPT sponsored response exactly as it reads on a hoarding.
Layered on top is the Ministry of AYUSH's advertising and self-regulation guidance, issued to registered practitioners, clinics, and manufacturers operating under the ayurveda, yoga and naturopathy, unani, siddha and homeopathy systems. This guidance extends the central Act's disease-list logic into practitioner conduct and digital advertising specifically, addressing formats the 1954 Act's drafters never anticipated — websites, social platforms, and by direct extension, AI-mediated conversational ads. It also sets expectations around identifying a formulation's manufacturing licence and a practitioner's registration when either is referenced in promotional material.
The third layer is the state Ayush directorate — a state-level regulatory body that registers AYUSH practitioners and clinics and, in several states, issues its own advertising-conduct rules that can be stricter than the central guidance. A clinic operating across Kerala, Karnataka, and Delhi is not clearing one rule set three times; it is clearing three potentially different rule sets once each, and a ChatGPT Ads campaign that targets users across state lines needs its claims checked against every relevant overlay before launch, not just the central Act and ministry guidance.
A fourth, quieter source of obligation sits inside the Drugs and Cosmetics Act, 1940 and its AYUSH-specific rules for manufacturing licences on ayurvedic, siddha and unani formulations. This isn't primarily an advertising statute, but it becomes one the moment a ChatGPT Ad names a specific product — the licence that legitimises the product's manufacture and sale is the same licence a follow-up question about "is this approved" needs to point back to. ICG treats this fourth layer as inseparable from the advertising-claims layers above it, because a compliant claim about a product that turns out to have no traceable licence is not actually a compliant campaign, it is a compliant sentence sitting on top of an unresolved regulatory gap.
Why AYUSH claims need root-level discipline, not surface edits
A static Google or Meta ad for an ayurveda brand is one fixed sentence, reviewed once, that stays fixed after approval. A ChatGPT Ad is a sponsored response inside a live conversation, and the assistant can be asked follow-up questions the original ad never anticipated — "does this actually cure diabetes," "how long until symptoms go away," "is this formulation government-approved." Each of those follow-ups pulls from whatever source material the brand has made available — landing page copy, structured data, product descriptions — and synthesises an answer that the user experiences as one continuous claim from the brand, not two separate assets reviewed on two separate days.
This matters more for AYUSH than for almost any other vertical because the underlying restriction is a specific, named disease list rather than a general "no unsubstantiated claims" standard. A general outcome claim like "improves overall wellness" is vague enough to survive most follow-up paraphrasing without tripping a specific violation. But a claim that even loosely associates with a scheduled disease — diabetes, cancer, sexual disorders — creates exposure the moment the assistant's synthesised answer names that disease alongside a treatment or cure framing, even if the original ad copy never used those exact words. The compliance unit, again, is the claim underneath the copy, and for AYUSH that claim has to be checked against a specific disease list, not just a general truthfulness standard.
Drug licence linkage compounds this. When an ayurveda ad references a named formulation — a proprietary churna, a branded oil, a specific tablet — a user's natural follow-up is "is this medicine safe" or "is this licensed," and the assistant will attempt to answer using whatever the brand has published about the product. If that source material doesn't clearly link the formulation to its manufacturing or marketing licence, the assistant's answer can either fabricate a plausible-sounding but unverifiable claim of approval, or surface silence that reads as evasion — neither of which serves the brand or the regulator's interest. ICG's practice is to make licence information part of the structured data feeding every AYUSH ChatGPT Ad, not an afterthought on a separate compliance page.
Panchakarma and other named protocol descriptions sit in a genuinely permitted zone that many brands under-use out of over-caution. Describing vamana, virechana, basti, nasya, and raktamokshana by name and process is factual service description, not a claim requiring substantiation — it is the pairing of that description with a disease-cure outcome that creates the violation. ICG's copy keeps the protocol detail, which is a strong trust and specificity signal for a ChatGPT Ad's synthesised answers, and removes only the outcome pairing.
What actually goes wrong, in anonymised real campaigns
The most frequent pattern ICG sees when an ayurveda or wellness brand runs its first ChatGPT Ads batch is direct disease-plus-cure pairing carried over from older print or social copy that was never re-audited against the 1954 Act's schedule. An anonymised ayurveda-clinic client's existing marketing line read "ayurvedic cure for diabetes, no side effects" — a construction that pairs a scheduled disease directly with a cure claim, the exact pattern the Drugs and Magic Remedies Act restricts. Carried into a ChatGPT Ad unchanged, the risk compounds because a user asking a natural follow-up ("does it really cure it") gets an assistant-synthesised answer built on that same unverified premise. The fix removed the disease name and cure verb entirely, replacing the line with a metabolic-wellness framing that describes the programme's approach without naming the scheduled condition or promising resolution.
A second recurring pattern involves an anonymised wellness-brand client whose copy leaned on testimonial language — "our patients say their PCOS is gone" — assuming that attributing the claim to a patient rather than the brand created distance from the restriction. It does not. A testimonial that states or implies a scheduled condition was cured carries the same exposure as the brand stating it directly, because the advertisement is still promoting the cure claim regardless of whose voice carries it. The fix kept testimonial language for service experience — consultation quality, staff attentiveness, facility comfort — and removed every outcome or cure attribution from quoted material.
A third pattern is drug-licence-specific: an anonymised ayurvedic-formulation manufacturer named a proprietary product in ad copy — "try our signature immunity blend" — without any licence reference in the source material feeding the campaign. When ICG ran the follow-up paraphrase test asking "is this product government approved," the assistant's synthesised answer, drawing from the brand's own thin product page, produced an ambiguous non-answer that read worse than either a clear yes or a clear no would have. The fix added the manufacturing licence number and issuing authority directly into the structured data behind the product page, so any follow-up question has a factual, checkable answer to draw from.
A fourth pattern is state-overlay specific: an anonymised multi-city wellness chain ran identical ChatGPT Ad copy across Kerala, Delhi, and Maharashtra locations, unaware that Kerala's Ayush directorate applies a stricter practitioner-advertising standard than the central guidance alone would require. Copy that cleared the central AYUSH review still needed a second pass against the state-specific rule before the Kerala-targeted ads could go live. A fifth, related pattern showed up in an anonymised yoga-and-naturopathy brand's campaign, where the ad copy referenced "clinically proven results" without any underlying clinical study to point to — a phrase borrowed from a competitor's website rather than from the brand's own substantiation, which meant the claim failed not because "clinically proven" is inherently prohibited but because the brand had nothing to produce when ICG asked for the study during the pre-push review. The fix was either sourcing an actual study reference or, more commonly, removing the phrase in favour of describing the practitioner's training and the protocol's traditional basis instead. The consistent lesson across all five patterns: strip disease-cure pairing regardless of who states it, link every named formulation to its licence in the source data, check state overlays before assuming central clearance is sufficient everywhere the campaign runs, and never let a claim survive the pre-push review if the brand can't produce its substantiation on request.
Worked examples, before and after
"Ayurvedic cure for diabetes — reverse your blood sugar naturally in 30 days, no side effects."
"Ayurvedic metabolic wellness programme, developed with AYUSH-registered practitioners. Explore our approach to diet, lifestyle and herbal support."
The rewrite removes the scheduled disease name, the cure verb, the timeline promise, and the unverifiable "no side effects" claim, replacing all four with a wellness-programme framing that describes the approach factually without naming the condition or promising an outcome.
"Panchakarma detox that eliminates cancer risk — thousands of patients trust our 21-day protocol."
"21-day panchakarma protocol — vamana, virechana and basti administered under registered vaidya supervision. Learn what each stage involves."
"Eliminates cancer risk" pairs a scheduled disease with a prevention claim, and "thousands of patients trust" is unverifiable social proof. The rewrite keeps the protocol name, duration, and supervision detail — all factual, all specific, all a strong trust signal for an assistant's synthesised follow-up answer — and drops only the disease association and the social-proof number.
"Our signature immunity blend — India's most trusted herbal formulation, government approved."
"Herbal wellness formulation, Licence No. [XXXXX] under the Drugs and Cosmetics Act. Ask about ingredients and usage."
"India's most trusted" is an unverifiable superiority claim and "government approved" without a specific licence reference is exactly the kind of vague assertion that produces an ambiguous or fabricated answer when an assistant is asked to substantiate it. The rewrite states the actual licence number, which is checkable and closes the exposure a follow-up question would otherwise open.
Pricing
Retainers from ₹20,000/month · Custom-scoped per engagement.
The realistic sequence, not the worst-case story
An AYUSH-related audit or complaint typically starts from one of three sources: a competitor or consumer complaint routed to the state Ayush directorate or a district drug-control authority, a routine review by OpenAI's ad partner network flagging a claim pattern, or a direct central-government action under the Drugs and Magic Remedies Act, which does provide for criminal penalties in its more serious provisions — a distinguishing feature versus the purely civil/administrative exposure under ASCI or NMC review for other healthcare verticals. None of these move instantly; a state directorate inquiry typically runs on a notice-response cycle measured in weeks, and platform-side reviews tend to pause the specific ad unit pending clarification rather than the whole account.
The realistic immediate consequence is a paused campaign, a notice requiring the specific claim's withdrawal, and a gap in attribution data — not an instant penalty. The disease-list restriction under the 1954 Act does carry statutory penalty provisions that are more serious on paper than ASCI's complaint-remedy process, which is precisely why ICG treats scheduled-disease scanning as a hard pre-push gate rather than a soft best-practice recommendation for this vertical specifically. The larger practical cost, as with any healthcare ChatGPT Ads pause, is time inside the 90-180 day first-mover window most Indian healthcare and wellness brands are still operating in on this ad surface — a paused AYUSH campaign loses conversation-history and completion-rate data that's expensive to rebuild once a competitor's clean campaign has kept running.
ICG's response protocol is documented and fast: the review trail from the pre-push checklist — including the scheduled-disease scan and licence-linkage records — is produced immediately on request, the flagged claim is withdrawn or rewritten within 24-48 hours, and the rest of the batch, which went through the same checklist, continues running unless the platform or directorate specifically requests a broader pause. Scale and Enterprise clients get a dedicated compliance point of contact for exactly this scenario; Starter and Growth clients route through the same review-trail process on request. The objective isn't zero risk in a regulatory environment with a named disease list and state-by-state overlays — it's making sure the answer to any inquiry is a documented review trail produced in hours.
AYUSH-clean ChatGPT Ads, built by a healthcare-only team
ICG's scheduled-disease scan and licence-linkage checklist run on every batch, every tier, before a single rupee of media spend goes live.
