SOURCED FROM META'S OWN POLICIES, SEPTEMBER 2026

What actually gets a WhatsApp Business account banned

Meta does not ban business accounts for using automation, and it does not ban them for sending too many messages. It acts on two things. User feedback signals, meaning blocks, reports, mutes, archives and low read rates. And policy category violations, meaning prohibited verticals, impersonation, spam, template misclassification and messaging without opt-in. In Meta's words, it may limit or remove access "if you receive significant amounts of negative feedback, cause harm to WhatsApp or our users, violate or encourage others to violate our terms or policies". Almost everything else circulating as WhatsApp ban lore is either operator practice or out of date.

This page quotes the policy text, names the source, and flags every place where Meta's own documents disagree with each other. Where we could not find a rule, we say so instead of inventing one.

Policy positions on this page were read from Meta's published documents and the TDRA's published policies on Thursday 3rd September 2026. Meta changes these documents without notice. Two of the facts below are scheduled to change on 1 October 2026 and are labelled where they appear.

The short answer

What Meta actually names as a trigger

Four documents govern a WhatsApp Business account, and they nest. Most confusion in this market comes from people arguing about one document while the answer sits in another. The generative AI clause everybody argued about in late 2025, for example, is not in the Messaging Policy at all. It is in the Business Solution Terms.

DocumentWhat it governsLast updated, as shown on the page
WhatsApp Business Messaging PolicyWhat you may say and to whom. Opt-in, templates, prohibited verticals, enforcement.No date shown on the page
WhatsApp Business Terms of ServiceThe contract. Contains the limit, throttle, suspend or terminate clause.16 February 2024
WhatsApp Business Solution TermsPlatform and API users. Data restrictions and the AI Providers clause.6 March 2026
WhatsApp Messaging GuidelinesHow Meta detects and reviews violations, automated and human.17 March 2025

Note for anyone chasing a broken link: the old WhatsApp Commerce Policy URL now redirects to the Business Messaging Policy. There is no longer a standalone commerce policy page at that address, and the Messaging Policy points to the Meta Commerce Policy instead. Several Meta documents still link to the retired URL.

The enforcement clause, verbatim

"We may limit or remove your access to or use of the WhatsApp Business Services if you receive significant amounts of negative feedback, cause harm to WhatsApp or our users, violate or encourage others to violate our terms or policies, or to maintain a high quality experience on WhatsApp, as determined by us in our sole discretion."

"Driving high quality conversations between people and businesses is a top priority. People can block or report businesses and our systems will limit the amount of messages a business can send or calls a business can initiate if the business' quality tier is low for a sustained period of time."

WhatsApp Business Messaging Policy, Section 7. whatsappbusiness.com/policy

The developer-facing enforcement page is more specific about what actually gets flagged, and it is worth reading the list slowly, because one item on it is the single most common deliberate cost-saving trick in this market.

"If Business Accounts repeatedly violate the WhatsApp Business Terms of Service, such as sending spam, template misclassifications, or high-risk policy categories such as adult content, sale of alcohol and tobacco, drugs, gambling and unsafe supplements, they might start seeing messaging restrictions that gradually increase in duration."

"In some cases, where there is evidence of a policy violation that causes severe harm to our users, such as child exploitation, scams, terrorism, or the sale of illegal drugs, WhatsApp will immediately offboard these Business accounts."

WhatsApp Business Platform policy and spam enforcement, last updated 21 May 2026. developers.facebook.com

Template misclassification is named by Meta as an enforcement trigger. That means dressing a marketing message as a Utility or Authentication template to pay the cheaper rate is not a clever workaround. It is on Meta's own published list of things that get accounts restricted.

Quality

The quality rating system as it works today, and the part of it that no longer exists

The correction most guides have not made

Meta's own help article carries this line: "The phone number statuses of Flagged and Restricted have been discontinued as of 7 October 2025." Any article built around "your number went Flagged, you have seven days to fix it before you drop a tier" is describing a system Meta switched off. If a vendor is still selling against that system, they have not read the source this year.

What exists today is a three-state phone number quality rating, calculated over a rolling seven days from customer feedback.

"Your quality rating is a rating based on the recent messages that your customers have received over the past seven days. This rating is determined by feedback from your customers, such as the reason why they blocked your number and other reporting issues."

Meta Business Help Centre. facebook.com/business/help/896873687365001
RatingMeta's definitionWhat it means for you
GreenHigh quality"You can become eligible for more features when your business maintains a Green quality rating."
YellowMedium qualityHover the rating in WhatsApp Manager to see block reasons, where Meta makes them available.
RedLow qualitySame visibility. Sustained low quality is what the Messaging Policy says drives message limiting.

The score is driven by what users choose when they block you. Meta's published block reasons are No longer needed, Didn't sign up, Spam, Offensive messages and No reason. "Didn't sign up" is the one that tells you your opt-in is broken rather than your copy.

What the score is actually calculated from

"Your message quality is based on how messages have been received by WhatsApp users over the past seven days and is weighted by recency. It is determined by a combination of user feedback signals like blocks, reports, mutes, archives, and reasons users provide when they block you."

"Numbers with high traffic commonly experience quality changes within short intervals (even within minutes)."

Cloud API send-messages documentation, last updated 21 May 2026. developers.facebook.com

Mutes and archives count against you. That is materially stricter than the usual advice that only blocks and reports matter. A customer who is too polite to block you but quietly archives the thread is still moving your rating. So is a recipient who simply never opens the message, because template quality separately reacts to low read rates.

Template quality is a second, separate score

Your phone number has a rating. Every message template has its own. Meta's four template states are GREEN for high quality with little or no negative feedback, YELLOW for medium quality where a template "has received negative feedback from multiple WhatsApp users, or low read-rates, and may soon become paused or disabled", RED for low quality and "in danger of being paused or disabled soon", and UNKNOWN where no feedback or read-rate data exists yet. All four can still be sent.

Instance at lowest qualityConsequence
1stTemplate paused for 3 hours
2ndTemplate paused for 6 hours
3rdTemplate disabled

Meta adds: "Pausing will initially not impact the business phone number from which the message template was sent... However, if a business consistently sends message templates that reach a Low quality status, the phone number may eventually be impacted." Template pausing documentation, last updated 17 June 2026.

Template pacing, or why your broadcast only half-sent

Before you conclude you have been restricted, check whether Meta simply paced you. This mechanism is barely covered in vendor content and it explains most "my blast stopped" panics.

"When a template is paced, messages will be sent normally until an unspecified threshold is reached. Once this threshold is reached, subsequent messages using that template will be held to allow enough time for customer feedback. Once Meta receives a good quality signal, subsequent messages using that template will be scaled to the entire target audience. If Meta receives a bad quality signal, subsequent messages using that template will be dropped."

Template pacing documentation, last updated 21 May 2026. developers.facebook.com

Held messages appear in the API as held_for_quality_assessment. Dropped ones return a failed status with code 132015. Pacing applies to marketing and utility templates, and catches new templates, recently unpaused templates, and any template without a GREEN rating. Meta states its goal as: "even if paced, campaign messages with highest throughput still get delivered within an hour (99 percentile)."

Volume

Messaging limits start at 250, not 1,000

What vendors publish

"Tier 1 gives you 1,000 messages per day. Warm up your number and you'll move to Tier 2."

What Meta documents

"Newly created business portfolios have a messaging limit of 250." The tiers are 250, then 2,000, then 10,000, then 100,000, then unlimited, and they are set at portfolio level, not number level.

A messaging limit is not a cap on messages. It is a cap on people.

"Messaging limits are the maximum number of unique WhatsApp user phone numbers your business can deliver messages to, outside of a customer service window, within a moving 24-hour period."

"Messaging limits are calculated and set at the business portfolio level and are shared by all business phone numbers within a portfolio. This means that if a business portfolio has multiple business phone numbers, it's possible for one number to consume all of the portfolio's messaging capability within a given period."

Messaging limits documentation, last updated 21 May 2026. developers.facebook.com

That second paragraph kills a common piece of advice. Splitting your sending across several numbers does not buy you more volume, because the limit is shared across the whole portfolio. It only isolates quality risk to one number. Anyone selling extra numbers as a way to send more is wrong on Meta's own text.

Getting from 250 to 2,000

There are three routes, and two of them are administrative rather than behavioural.

RouteWhat it requiresHow long Meta says it takes
Meta business verificationVerify your business directly with Meta"A verification decision can be made in as little as 10 minutes and as long as 14 working days."
Partner-led verificationYour onboarding partner verifies you"Most requests are resolved within minutes, but some can take up to 24 hours."
Quality messaging pathDeliver messages outside customer service windows to unique users, using templates with a high quality ratingOver a 30-day period. See the wording conflict below.

Meta's own documents disagree here, so we report both

The developer documentation describes the quality messaging path as "Send 2,000 delivered messages outside of customer service windows to unique WhatsApp user phone numbers within a 30-day moving period, using templates with a high quality rating." The Meta Business Help Centre article on the same scaling path says it "requires a minimum of 1,000 conversations over 30 consecutive days." Both are Meta. We are not going to pick one and present it as settled.

Once you are past 2,000, the ladder is automatic. Meta's criteria are that "you are sending high-quality messages across all of your business phone numbers and templates" and that "in the last 7 days, your business has utilized at least half of your current messaging limit". If both are met, "your portfolio's limit increases by one level within 6 hours". Denials arrive as an account_alerts webhook with an alert type of eligibility deferred, failed, or need more info.

An honest gap: there is no published downgrade rule

The Messaging Policy says Meta's systems "will limit the amount of messages a business can send... if the business' quality tier is low for a sustained period of time". But since Flagged and Restricted were discontinued on 7 October 2025, Meta's current published documentation no longer states a mechanism, a threshold or a duration for a messaging-limit downgrade. The messaging-limits page describes increases only. We cannot source a current downgrade rule, so we are not going to invent one, and neither should anyone quoting a specific number of days at you.

The ceiling nobody sees: per-user marketing limits

Separately from your account limit, Meta caps how many marketing templates an individual user receives from any business, based on "a dynamic view of an individual's recent marketing message read rate and how many messages they currently have in their inbox from friends, family, and businesses".

  • Every delivered marketing template counts towards that user's limit. If the user replies, the 24-hour window opens and messages inside it do not count.
  • Marketing templates are not currently delivered to United States phone numbers at all.
  • Per-user limits are not active for senders or recipients in the European Economic Area, United Kingdom, Japan or South Korea.
  • The UAE is on neither exclusion list. Per-user marketing limits apply to UAE senders and UAE recipients.
  • Hitting the limit returns error code 131049, and Meta advises waiting at least 24 hours before retrying that user.

There is also a user-controlled switch that bypasses your CRM entirely. WhatsApp gives users an Offers and announcements setting to stop marketing from your business. If a user has stopped them, "the API will process the request but not send the message", returning code 131050. Subscribe to the user_preferences webhook if you want to see stops and resumes. Interested and Not interested feedback does not fire that webhook.

Consent

Opt-in: the rule most businesses get wrong in both directions

Some businesses message people who never gave them a number, which is a straightforward violation. Others refuse to message anyone unless they ticked a WhatsApp-specific box, which is stricter than Meta requires and costs them conversations they were entitled to have.

"You may only contact people on WhatsApp if: (a) they have given you their mobile phone number; and (b) you have received opt-in permission from the recipient confirming that they wish to receive subsequent messages or calls from you."

"You must respect all requests (either on or off WhatsApp) by a person to block, discontinue, or otherwise opt out of communications from you via WhatsApp, including removing that person from your contacts list. Do not confuse, deceive, defraud, mislead, spam, or surprise people with your communications."

WhatsApp Business Messaging Policy, Section 1. whatsappbusiness.com/policy

Note "either on or off WhatsApp". If someone tells your receptionist to stop messaging them, that counts, and your CRM has to carry it.

The clarification almost nobody publishes

"As per the November 2024 WhatsApp Business Messaging Policy update, before messaging people on WhatsApp, businesses are required to obtain opt-in permission, which can be general and not specifically for WhatsApp, as long as businesses comply with all local laws."

Getting opt-in documentation, last updated 16 June 2026. developers.facebook.com

Meta's three stated requirements are that businesses clearly state the person is opting in to receive communication, clearly state the business name they are opting in to hear from, and comply with applicable law. Meta's accepted methods are SMS, a website, a phone IVR flow, and in person or on paper. Meta also advises businesses to "monitor quality rating, especially when rolling out new opt-in methods", which is a quiet admission that a bad opt-in shows up as a quality problem before it shows up as a violation.

UAE businesses: Meta's opt-in rule is not the only opt-in rule

Cabinet Resolution No. 56 of 2024 on Telemarketing, in force 27 August 2024, defines telemarketing to include "marketing text messages and marketing messages through social media applications". It applies to all companies licensed in the State, including those in free zones. Article 4 requires a company to "create a communication channel for Consumers interested in obtaining marketing information, and marketing communication is only made with these Consumers", to use local numbers registered under its commercial licence, and not to contact consumers listed on the Do Not Connect Register supervised by the TDRA.

Report this honestly: the definition in Article 1 clearly sweeps in social-media marketing messages, but the operative controls in Articles 4 and 5, including the 9:00 am to 6:00 pm window, call recording and callback frequency, are drafted in the language of phone calls. How they map onto a WhatsApp template message is not specified in the text. Treat the opt-in channel and the DNCR as applying. Get a lawyer's view before you treat the timing window as settled either way.

The AI clause

What Meta's AI rule actually says, and what vendors claim it says

In October 2025 a wave of content announced that Meta had banned AI chatbots from WhatsApp. It had not. Here is the actual clause, which lives in the Business Solution Terms and not in the Messaging Policy, quoted at the length that matters.

"AI Providers. Providers and developers of artificial intelligence or machine learning technologies, including but not limited to large language models, generative artificial intelligence platforms, general-purpose artificial intelligence assistants, or similar technologies as determined by Meta in its sole discretion ("AI Providers"), are strictly prohibited from accessing or using the WhatsApp Business Solution, whether directly or indirectly, for the purposes of providing, delivering, offering, selling, or otherwise making available such technologies when such technologies are the primary (rather than incidental or ancillary) functionality being made available for use, as determined by Meta in its sole discretion; provided, however, that such technologies may be made available to WhatsApp users who have registered phone numbers with a European Economic Area or Brazil country code. Notwithstanding the foregoing, you may retain an AI Provider as your Third Party Service Provider in accordance with these WhatsApp Business Solution Terms."

WhatsApp Business Solution Terms, last modified 6 March 2026. whatsapp.com/legal/business-solution-terms

Read it strictly and three things fall out.

What is prohibited

An AI provider using the WhatsApp Business Solution to distribute its own assistant where that assistant is the primary functionality on offer. This is the general-purpose assistant distribution case.

What is explicitly permitted

"You may retain an AI Provider as your Third Party Service Provider." A business running an AI assistant for its own customers is a customer of an AI provider, not an AI provider.

The carve-out nobody reports

The prohibition does not apply to users with an EEA or Brazil country code. General-purpose assistants may still be offered to those numbers.

Meta's own transition banner, captured in November 2025, gave the dates: "We're updating the WhatsApp Business Solution Terms as of October 15, 2025... These changes will go into effect: On the day you sign up to use our services on or after October 15, 2025. On January 15, 2026, if you signed up to use our services before October 15, 2025."

The vendor framing to ignore

Several widely-shared vendor articles, including one from respond.io published on 27 October 2025, describe the rule as banning "open-ended" chatbots while permitting "structured" ones. That distinction does not appear anywhere in the terms. The line Meta draws is about who is offering the technology and whether the AI itself is the product, not about how conversational the bot is. Nothing in the clause caps how capable your assistant may be, and nothing in it requires your assistant to be menu-driven.

The rule that does constrain your AI, which is older and separate

"You may use automation when responding during the 24-hour window, but must also have available prompt, clear, and direct escalation paths. These escalation paths include: In-Chat Human Agent transfer / Phone number / Email / Web support (on the business website) / In-store visits (ex: visit retail store, bank branch), or / Support form"

WhatsApp Business Messaging Policy, Section 2. whatsappbusiness.com/policy

That is the operative rule for a business AI assistant, and it is a design requirement rather than a prohibition. There must be a fast, obvious way to reach a human. There is also a data constraint that bites AI vendors rather than their clients: under the same Solution Terms you may fine-tune on Business Solution Data only for a model that is "for your exclusive use". Pooling client conversation data to train one shared model is prohibited, and Meta says it "may terminate your account and revoke your access" for breaching those restrictions.

One more line from Meta's own guidance is worth pinning up, because it turns bot quality from a nice-to-have into a compliance input: Meta names, among the factors influencing a number's quality rating, "the business's overall reputation, the bot functionality and how quickly questions and other enquiries are resolved". A bad bot is not just a bad experience. It is an input to the score that governs how many people you may message.

Category risk

The prohibited verticals, and why they matter more in the UAE

This is the most under-reported section of the Messaging Policy for a Gulf audience, because several of the categories on it are ordinary, licensed, respectable businesses here.

"We prohibit use of the WhatsApp Business Services for buying, selling, promoting, or otherwise facilitating the exchange of certain regulated or restricted goods and services listed below, as determined in our sole discretion."

WhatsApp Business Messaging Policy, Section 4.

The list, in Meta's own order: firearms; alcohol and tobacco; drugs, whether prescription, recreational or otherwise; medical and healthcare products; endangered species; live non-endangered animals excluding livestock; hazardous goods and materials; real, virtual or fake currency, including ICOs and binary options; body parts or fluids; business models or services Meta determines may be fraudulent, misleading, offensive, deceptive, exploitative, inappropriate or exerting undue pressure on targeted groups; real money gambling and gaming; adult products and services; dating services; multi-level marketing; and payday loans, paycheck advances, peer-to-peer lending, debt collection and bail bonds.

The sentence that catches licensed UAE businesses

"These prohibitions apply irrespective of the global or local licenses, registrations, or other approvals your business may hold."

A DHA-licensed aesthetic clinic, a licensed debt collection firm and a licensed matchmaking service are all touching this list, and Meta states plainly that a local licence does not cure it. Meta's regulated-vertical exceptions are country-gated, and the only allowed-country tables published are for Online Gambling and Gaming, Over-the-Counter Drugs, and Alcohol. The UAE appears on none of them. There is no allowed-country list at all for medical and healthcare products.

Two further absolutes worth knowing. On the free WhatsApp Business app, Meta writes: "You are prohibited from messaging about any Regulated Verticals on the WhatsApp Business App." And on the Platform, messaging about a regulated vertical "may not be sent to people under 18 years of age".

Politics and public bodies are handled separately and bluntly. Meta prohibits use of the Platform by "Political Parties, Politicians, Political Candidates, and Political Campaigns" and by "Law Enforcement Agencies, Military Services, National Security and Intelligence Agencies". Government entities are permitted but "require access through a Solution Provider".

What this means for a Dubai clinic, stated carefully

Meta's prohibition is written around products, and it is enforced "as determined in our sole discretion" rather than adjudicated. A clinic confirming an appointment a patient booked is not plausibly promoting a medical or healthcare product. A clinic broadcasting a promotion for an IV drip package or a prescription weight-loss programme plausibly is. Meta has not drawn that line itself, and we are not going to pretend it has. What Meta has written down is this:

  • "Don't use WhatsApp for telemedicine or to send or request any health related information, if applicable regulations prohibit distribution of such information to systems that do not meet heightened requirements to handle health related information." (Business Messaging Policy, Section 3)
  • "We make no representations or warranties that our Business Services meet the needs of entities regulated by laws and regulations with heightened confidentiality requirements for personal data, such as healthcare, financial, or legal services entities." (Business Terms of Service, Section 4, last updated 16 February 2024)

On the local side, be precise. HIPAA is United States law and a Dubai clinic treating Dubai patients is not a HIPAA covered entity. Any page telling a UAE clinic it must be HIPAA compliant as a matter of law is wrong. The UAE analogue is Federal Law No. 2 of 2019 on the Use of Information and Communication Technology in Health Fields, with its Executive Regulation and Decision No. 51 of 2021, layered with the emirate health authority's own rules. Note also that the widely-quoted penalty band of AED 500,000 to AED 700,000 is not a general health-data-breach fine: it is the penalty attached under Article 24 to breaches of Article 13, the prohibition on storing, processing, generating or transforming health data outside the State. It is not the penalty for messaging a patient. HIPAA does return in Dubai by a side door, as a procurement standard rather than a law, because the DHA Standards for Telehealth Services require telehealth platforms to hold HIPAA compliance certification. That is a reason to keep clinical conversations off WhatsApp, not a reason to claim WhatsApp is illegal here.

Surface choice

The free app is the riskier surface, not the API

Most prospects assume the WhatsApp Business Platform, meaning the API, is the dangerous route and the free WhatsApp Business app is the safe one. On Meta's own documented behaviour, the opposite is true.

WhatsApp Business appWhatsApp Business Platform (Cloud API)
Governed byMessaging Policy, Business Terms, Messaging GuidelinesThe same, plus the Business Solution Terms
Enforcement visibilityAn in-app ban noticeWarnings, then graduated blocks, itemised in Business Support Home, plus an account_update webhook
Graduated restrictionsNot documented. A self-clearing restricted state, then a banDocumented ladder: 1 or 3 day template block, then 5, 7 or 30 day full block, then account lock, then permanent disable
AppealRequest review in the app. If the option is absent, the ban is finalRequest Review in Business Support Home. All violations appealable within 90 days. Outcome Unchanged or Reversed
Quality ratingNot exposedGreen, Yellow, Red over 7 days, visible in WhatsApp Manager with block reasons
Regulated verticalsEntirely prohibitedPermitted only in named countries, with licence and age gating. The UAE is on none of the lists
AutomationConsumer rule applies: "Don't bulk message, auto-message, or auto-dial using WhatsApp"Automation explicitly permitted in the 24-hour window, with escalation paths

On the app you get one binary outcome and possibly no appeal at all. Meta's consumer-side guidance is explicit that "WhatsApp uses both machine learning technology and reports from users to detect and ban accounts that send unwanted automated messages", and that "frequent use of broadcast messages may lead people to report your messages, and we ban accounts that are reported multiple times". Broadcast lists also only reach people who have already saved your number, so their real reach is far smaller than the list size suggests.

On the Platform you get a warning, a named policy, an itemised violation, a 90-day appeal window and a documented ladder before anything becomes permanent. For a UAE business running any real outbound volume, the API is the lower ban-risk surface.

Recovery

What happens when you are restricted, and what an appeal really does

The Platform escalation ladder, in Meta's own order:

  • "Initially, WhatsApp Business Accounts will get a warning with information on the policy they violated."
  • "1 or 3 day block on sending marketing, utility, and authentication template messages and adding additional phone numbers to the account"
  • "5, 7, or 30 day block on sending any messages and adding additional phone numbers to the account"
  • "An account lock, which is an indefinite block on sending any messages; can only be removed via an appeal"
  • "Eventually be permanently disabled from the WhatsApp Business Platform, if the business does not make changes after multiple warnings and feature limits or blocks"

The appeal itself is a short form. From Business Support Home, select the WhatsApp Business Account, choose the violation, click Request Review, enter supporting details and submit. The request moves to the In Review tab and the outcome is either Unchanged or Reversed. Meta's Business Help Centre adds that "all policy violations can be appealed within 90 days of being received", and that restricted or disabled accounts can still appeal, with the account returning to its previous status if the issue is reversed.

Two things to set expectations honestly

Timing is inconsistent across Meta's own pages. The developer documentation says an appeal decision "typically takes 24 to 48 hours". The Business Help Centre says reviews "typically take 24 hours". Both are Meta, and we are reporting the disagreement rather than resolving it.

Not everything is appealable. Meta writes: "Not all spam violations might be appealed. Businesses might have to wait until the restriction period ends to begin messaging again." And separately: "Violations might be appealed or acknowledged, based on violation type and eligibility."

On the free app, the mechanism is the consumer ban system and it is far less forgiving. A banned account shows "This phone number can't use WhatsApp", with a Request review option. Meta's own line is the one to plan around: "If you don't see an option to request a review, it means the ban is final and can't be appealed." One phone number is reviewed per appeal. You cannot access chat history or backups while banned. A banned account automatically loses its Meta Verified subscription, and the subscription is cancelled without a refund. Meta also states that "third party services can't ban your WhatsApp Business account or remove your account ban", which is worth remembering the next time somebody offers to unban you for a fee.

There is a lighter state on the app side that clears itself. If you see "Your account is restricted right now" or "Your account may be restricted soon", Meta says no action is needed and "the restriction automatically ends once the allocated time passes". While restricted, you can keep talking to contacts and groups you have exchanged messages with before, but you cannot start new chats, create groups, add members or use linked devices.

How the decision gets made at all: Meta says automated processing "is central to our review process and automates decisions for certain areas where account behavior or reported message content is highly likely to be in violation", and that "when an account, group or community requires further review, our automated systems send it to a human review team to make the final decision".

What we will not tell you

How many reports it takes to trigger a ban. Meta publishes no number. Nor does it publish a success rate for appeals or a recovery rate for banned accounts. Any vendor quoting you either figure has made it up, and you should treat everything else they told you with the same suspicion.

The grey zone

Safe, consequential, and what operators actually do

Most content on this subject blends Meta's rules with market habits until you cannot tell which is which. Here they are separated. The middle column is policy with the consequence named. The bottom section is operator practice, labelled as operator practice, because that is what it is.

MoveStatusThe consequence, named
Replying to an inbound message with automation inside the 24-hour windowSafeExplicitly permitted, provided you offer prompt, clear escalation paths to a human.
Messaging a customer who gave you their number and opted in generally, not specifically for WhatsAppSafe under Meta's ruleMeta's opt-in doc permits general opt-in. UAE marketing rules still apply on top.
Running a capable AI assistant for your own customersSafeThe AI Providers clause targets distributing an assistant as the product, and permits retaining an AI provider as your service provider.
Sending a marketing message under a Utility or Authentication templateTriggers enforcementMeta names "template misclassifications" on its enforcement list. Outcome path: warning, then a 1 or 3 day template block, escalating.
Messaging bought or scraped numbersViolationFails the opt-in rule outright. Drives "Didn't sign up" blocks, which is the block reason that moves quality fastest. In the UAE it also runs into Cabinet Resolution 56 of 2024 and the DNCR.
Promoting regulated goods from a UAE numberViolationThe UAE is on no allowed-country list. Meta says a local licence does not cure it. High-risk categories are named on the enforcement page as a driver of escalating restrictions.
Sending clinical results or advice over WhatsAppDo notMeta's policy says do not use WhatsApp for telemedicine or health information where regulations demand heightened handling, and its Terms decline to warrant fitness for healthcare entities.
Sending a high volume of unread marketing templatesDegrades you without any blockLow read rate alone can push a template to YELLOW or RED. Three low-quality instances disables the template.

Why template misclassification is so tempting, in numbers

It is worth seeing the price gap that makes template misclassification tempting in the first place. These are Meta's own published UAE rates, pulled from Meta's rate system on 3 September 2026. Meta publishes AED directly, so these are not currency conversions.

CategoryUSD per messageAED per message
Marketing (one flat rate, no volume tiers in the UAE)0.04990.1832
Utility (first tier, up to 100,000 per month)0.01570.0576
Authentication (first tier, up to 300,000 per month)0.01570.0576
Service0 today0 today

Rates as at 3 September 2026, from Meta's own pricing system. Meta raised the UAE marketing rate on 1 October 2025, which is why stale vendor pages still show $0.0385. Two changes are scheduled for 1 October 2026: service messages become chargeable, and utility templates delivered inside the 24-hour window stop being free. Meta's main pricing page and its upcoming-changes page currently contradict each other on this, so verify both before you budget.

A marketing template costs roughly three times a utility template in the UAE. That gap is exactly why operators reclassify, and it is exactly why Meta named the practice on its enforcement page. The saving is real. So is the escalating restriction.

Operator practice, labelled as such

None of the following is Meta policy. It is what people in this market actually do, with the real risk stated.

Number warm-up schedules

Vendors publish ramps such as 10 to 20 messages a day in week one rising to 80 to 100 by week three. No Meta document prescribes a warm-up ramp. The only Meta-documented ramps are the messaging-limit ladder and template pacing. The risk in following a made-up schedule is not that it harms you, it is that it gives false confidence in place of watching your actual quality rating.

Batching sends of about 500 at a time

Operators describe staggering batches of roughly 500 on a number rated high quality, on the reasoning that Meta's enforcement is a black box. It is defensible, since quality is computed from seven days of feedback and pacing throttles you anyway. It is not a documented rule, and it does not raise your portfolio limit.

Splitting sending across several numbers

This isolates quality risk to one number, which is real. It does not buy volume, because messaging limits are portfolio-wide and shared. Anyone selling extra numbers as extra capacity is contradicting Meta's published text.

Suppressing non-responders early

Sensible against per-user marketing limits and read-rate-driven template quality, both of which are real mechanisms. But Meta publishes no threshold, so any specific rule such as "suppress after two no-replies" is a judgement call, not compliance.

FAQ

Questions people actually ask

Does using an AI chatbot get your WhatsApp Business account banned?

No. The WhatsApp Business Messaging Policy states: "You may use automation when responding during the 24-hour window, but must also have available prompt, clear, and direct escalation paths." Meta lists those paths as in-chat human agent transfer, phone number, email, web support, in-store visits or a support form.

A business running an AI assistant to serve its own customers is inside the rules. What Meta acts on is the outcome of the messaging, meaning blocks, reports, mutes and low read rates, not the fact that a machine wrote the reply.

Did Meta ban AI chatbots on WhatsApp in October 2025?

No. The clause added to the WhatsApp Business Solution Terms is headed "AI Providers" and it bars providers and developers of AI technologies from using the WhatsApp Business Solution to deliver those technologies "when such technologies are the primary (rather than incidental or ancillary) functionality being made available". The same clause then says explicitly: "you may retain an AI Provider as your Third Party Service Provider".

It targets the distribution of general-purpose assistants as a product, not a business running an assistant for its own customers. Meta's own transition banner dated the updated Terms 15 October 2025, taking effect on sign-up from that date, and on 15 January 2026 for accounts that signed up before it.

How many messages can a new WhatsApp Business API account send per day?

250. Meta's developer documentation states that newly created business portfolios have a messaging limit of 250, meaning 250 unique WhatsApp user phone numbers you can message outside a customer service window in a moving 24-hour period. The limit is set at business portfolio level and shared by every phone number in that portfolio.

It rises to 2,000, then 10,000, then 100,000, then unlimited. The widely repeated figure of 1,000 messages per day for Tier 1 is out of date.

What does a red quality rating actually do to your account?

Red is the lowest of three phone number states, Green, Yellow and Red, calculated from the last seven days of customer feedback. A template that reaches the lowest quality rating is paused automatically for 3 hours on the first instance, 6 hours on the second, and disabled on the third.

Meta says pausing initially does not affect the phone number, but that a number sending consistently low-quality templates may eventually be impacted. Note that the phone number statuses Flagged and Restricted were discontinued on 7 October 2025, so any guide describing a seven-day drop from Flagged is describing a system that no longer exists.

Can you appeal a WhatsApp Business account restriction, and does it work?

On the WhatsApp Business Platform, yes, in most cases. You select the violation in Business Support Home and click Request Review. Meta's Business Help Centre says all policy violations can be appealed within 90 days of being received. The outcome is either Unchanged or Reversed.

Meta's developer documentation says a decision typically takes 24 to 48 hours while its Business Help Centre says reviews typically take 24 hours. Meta also warns that not all spam violations can be appealed and that some businesses have to wait for the restriction period to end.

On the free WhatsApp Business app the position is harsher: if there is no Request review option in the app, the ban is final. Meta publishes no success rate for appeals, so nobody can honestly quote one.

Do you need a WhatsApp-specific opt-in to message customers?

No. Meta's own opt-in documentation says that under the November 2024 policy update, businesses must obtain opt-in permission "which can be general and not specifically for WhatsApp, as long as businesses comply with all local laws".

Three things must be true: you state clearly that the person is opting in to receive communication, you name the business they are opting in to hear from, and you comply with applicable law. Meta accepts SMS, a website, an IVR phone flow, or a signed paper form. In the UAE, Cabinet Resolution No. 56 of 2024 adds its own requirements on top for marketing messages.

Why did my WhatsApp broadcast only send to half the list?

Usually template pacing, not a ban. Meta holds messages from a newly created, recently unpaused, or non-Green template once an unspecified threshold is reached, waits for early customer feedback, then either scales the send to the full audience or drops the held messages. Held messages appear as message_status held_for_quality_assessment, and dropped ones return a failed status with code 132015.

A second cause is the per-user marketing limit, which stops marketing templates reaching individual users who are unlikely to engage; that returns error code 131049 and Meta advises waiting at least 24 hours before retrying.

Is WhatsApp banned in the UAE?

WhatsApp messaging works normally in the UAE. WhatsApp voice and video calling does not. VoIP is a regulated activity under the TDRA's Regulatory Policy on Voice over Internet Protocol, and TDRA's published list of permitted VoIP applications does not include WhatsApp. UAE press reported in February 2026 that WhatsApp calling remained restricted on mobile and desktop.

Nothing in that regime restricts WhatsApp text messaging, which is why messaging is the channel UAE businesses can build on. Voice is a separate question. A business call placed over WhatsApp is still WhatsApp voice traffic, so it sits inside the same VoIP regime, and we could find no source, primary or secondary, confirming either way whether a WhatsApp Business Calling API call connects to a consumer on a UAE mobile network. Treat that as unresolved rather than available.

Can a Dubai clinic use the WhatsApp Business Platform?

Carefully, and with legal advice. Meta's prohibited list includes "Medical and healthcare products", and the policy says those prohibitions apply "irrespective of the global or local licenses, registrations, or other approvals your business may hold". Meta's regulated-vertical exceptions are country-gated and the UAE appears on none of the published allowed-country lists.

Meta's policy also says do not use WhatsApp for telemedicine or to send or request health related information where regulations demand heightened handling, and its Business Terms decline to warrant that its services meet the needs of healthcare entities.

HIPAA is United States law and does not apply in the UAE; the UAE analogue is Federal Law No. 2 of 2019 on the Use of Information and Communication Technology in Health Fields, layered with the emirate health authority's rules.

Where we come in

How Learnmind builds this so it does not become your problem

We build AI customer communication for UAE businesses on the WhatsApp Business Platform, and the quality mechanics above are the constraint we design around rather than a footnote we bolt on afterwards. Meta names bot functionality and resolution speed as inputs to your number's quality rating, which means the assistant's job is not only to answer, it is to keep your rating green.

Two things we can point at publicly. For Alcaz Media, a UAE performance marketing agency, the published outcome is that "every lead gets a response within seconds. Whether the message comes in at 9am on a Tuesday or 2am on a Saturday, the prospect gets immediate, intelligent engagement," and that qualified leads land in the CRM with full contact details and conversation history. For Lola's Lashes, a UK beauty brand, co-founder Fraser Angus put it this way: "Just having it in the back of your head that 100,000 people saw a comment that said I was burning after they applied your glue... it's good to have the peace of mind knowing that there's an actual system that's taking care of it." Both are set out in full, in their own words, on our results page.

We also work with clients in the healthcare and aesthetics space. Due to the sensitive nature of these engagements, we are unable to publicly disclose their identities. If you are a healthcare provider and would like to learn more about our work in this sector, get in touch and we can share relevant details privately.

One pattern worth naming, because it comes up on almost every sales call we run in the Gulf and it is not in anyone's marketing: buyers here accept AI on text and reject it on the phone, and they say why. One Dubai owner overruled his own team, who wanted phone AI, with the line "AI can talk logical, but it cannot talk emotional." That is an observation from our own conversations, not a study, and we present it as such. It also happens to line up neatly with the regulatory picture, since messaging is the channel that works normally here and calling is not.

Have someone read your setup before Meta does

Fifteen minutes. We will look at your opt-in wording, your template categories and your quality rating, and tell you which of the risks on this page you are actually carrying. No slides.

Primary documents quoted on this page

This page is a summary of published policy and regulation, written for UAE business owners. It is not legal advice. Meta amends these documents without notice, and two of the pricing facts above change on 1 October 2026. Check the primary sources before you make a decision that depends on them.