Meta banned open-ended chatbots in January. Is your Arabic bot policy-safe?
Meta's January 15, 2026 policy requires AI on WhatsApp to perform concrete business tasks. No open-ended chat. Most Arabic WhatsApp bots built between 2023 and 2025 — when "ChatGPT for WhatsApp" was the deployment template — are exposed. We audit your bot for policy-risky flows, generate a task-specific re-architecture where needed, issue the TaskSpec Verified badge and compliance documentation for Meta reviews, and monitor for drift.
What's at stake
- Meta enforces reactively but consistently. Companion / entertainment AI hit the first enforcement wave in Feb–March 2026. Open-ended business bots are next as customer complaints accumulate.
- The 2023–2024 Arabic bot wave is over-exposed. Most MENA bots from that period were "wrap an LLM in a WhatsApp gateway and let it answer anything." That architecture is the failing pattern.
- Enforcement actions hit the WABA, not the bot. When Meta acts, your entire account gets the notification. The merchant has to identify the offending flow and fix it within 7–14 days.
- The documentation requirement is real. Meta wants written task-scope documentation on re-application. Bots that previously passed without scrutiny are being re-reviewed under the new rules.
What we do
A four-stage compliance audit:
- 1. Scope statement. We define what your bot is supposed to do, in writing, one paragraph. This is the artifact Meta wants. Most non-compliant bots fail here because no scope statement exists.
- 2. Flow inventory. Every conversational flow gets mapped — trigger, path, exit. Flows without clear triggers or exits are flows operating open-ended.
- 3. Boundary test. For each flow, what does the bot do when a user request falls outside scope? Three acceptable answers (handoff to human, fallback content, polite refusal). One unacceptable answer (the LLM freelances).
- 4. Prompt and guardrail review. For LLM-backed flows, we review the system prompt for explicit scope constraints and fallback behaviour. Arabic deployments need Arabic-native prompts, not translations.
Passing bots get the TaskSpec Verified badge + compliance documentation package. Failing bots get a remediation plan with specific architecture changes.
Pricing
| Tier | Price | What's included |
|---|---|---|
| Audit | $299 one-time | Full four-stage audit, remediation plan if needed, TaskSpec Verified badge if passing. Audit fee credits toward implementation if remediation is required. |
| Badge + Monitoring | $99/month | Audit + ongoing automated drift monitoring, monthly re-audit summary, badge maintenance. Includes one re-audit per quarter at no extra cost. 7-day free trial. |
| Enterprise | $499/month for 5+ bots | Audit + monitoring across a bot portfolio, quarterly re-audit cycle, priority remediation support, policy-change watch (we flag Meta policy updates within 48 hours of publication). 7-day free trial. |
Why us
- Vendor-agnostic. We audit bots on any stack — Wittify, Tactful, Intella, Teammates, or custom builds.
- Arabic-native prompt review. System prompts authored in Arabic produce noticeably more consistent scope adherence than translated prompts.
- Documentation Meta actually wants. Our compliance package is structured to match the artifacts Meta's reviewers ask for on re-application.
- Drift monitoring catches issues before Meta does. Monthly automated checks against your flow patterns flag scope creep early.
Frequently asked questions
Is Meta actually enforcing this policy?
Yes, reactively. A first wave hit companion/entertainment AI in Feb–March 2026. A slower customer-complaint-triggered wave is targeting open-ended business bots. Enforcement pace is increasing.
What happens if my bot gets actioned by Meta?
WABA notification with 7–14 day window. Required removal or rewriting of the offending flow. Documentation requirement on re-application. First-time actions typically resolve within 2–4 weeks if the merchant moves promptly.
Does the TaskSpec Verified badge mean Meta approved my bot?
No. The badge is third-party audit attestation. Meta doesn't run pre-approval for AI bots. The badge value is preparation — you have the documentation Meta will ask for, you've caught policy issues before they cost trust, and you've signalled responsibility to customers.
My bot is in Arabic — does the audit work?
Yes. The audit team works in Arabic and English. System-prompt review and scope-statement authoring happen in the language the bot is deployed in.
Can I just rewrite my system prompt and skip the audit?
You can pass on your own. The audit's value is catching gaps you didn't think of — fallback flow gaps, documentation gaps. Self-rewrites typically miss both.
How does this work with my bot vendor?
The audit is vendor-agnostic. We work with whatever stack you're on. Most MENA vendors expose enough flow architecture and prompt configuration to make audit straightforward.
What if Meta's policy changes again?
Subscription tiers include policy-change monitoring. Policy changes get flagged within 48 hours, affected bots in your portfolio identified, remediation queued if needed.
Run the free task-scope check.
5 minutes. We'll tell you which of your bot's flows are in the policy-risky band and what remediation would look like. Most audits cost $299; the check is free.
Run the free task-scope checkSix products. One studio.
Each product solves a specific Meta change. Pick the next one that applies to your stack.