AI Chat Compliance — Article 50(1) + (2) for Customer-Service AI.
Chatbots, AI sales agents, WhatsApp AI, email AI — Article 50(1) interaction disclosure and Article 50(2) synthetic content marking, on the same evidence chain as your voice agents. One signed receipt for every channel.
Article 50(1) vs Article 50(2)
Article 50(1) — Interaction disclosure
The user must be told they are talking to AI. Audact enforces it on every chat session, every channel, every locale — and logs that it happened.
Article 50(2) — Synthetic content marking
Generated text, images and audio must be machine-readable as AI-generated. Audact attaches cryptographic provenance (C2PA) to each one.
Wherever AI speaks for your brand, the obligation follows
Voice is just one channel. The same Article 50 disclosure rules — and the same evidence chain — cover every text channel your clients run:
At launch (Q4 2026)
EU-hosted only. Frankfurt (Ireland backup region). No US data path.
Integration patterns
Webhook fires on each outbound AI message. Audact attaches an Article 50(1) disclosure ID + Article 50(2) C2PA manifest reference.
Webhook fires on each inbound user message. Audact logs interaction-disclosure state, locale, and policy version for the session.
Put every channel on one evidence chain.One evidence chain by design. Voice, chat, email, WhatsApp — same Merkle root, same signed receipt a DPO can verify. Channel Continuity is included in the Audact platform (Agency Pro & Enterprise). Usage-based pricing for the standalone API is confirmed at the Q4 2026 launch — talk to us to scope it. Talk to sales On the roadmap
Chat compliance ships in Q4 2026 — the same governance and disclosure layer extended to web chat, WhatsApp and email. Talk to us to be first in line.
- Same evidence chain as voice — export a signed receipt in under a minute
- REST + webhook, no SDK lock-in
- Locale-aware disclosure copy out of the box