W3TURN AI Office — AI Usage Notice
Document owner Weturn Artificial Intelligence Developing Services L.L.C Effective date [Insert effective date] Draft last updated 6 August 2026 (v3) Status Draft for qualified legal review. Not yet published.
Version 3 note. Re-verified against
main@bb11bb6on 6 August 2026. Corrects the description of how agent knowledge is stored and retrieved, completes the provider list, and adds the voice channel. Every correction is itemised, with the contradicting source file named, indocs/legal/legal-pack-v2-code-review.md.
1. You are using AI systems
W3TURN AI Office uses artificial intelligence to help business users plan work, select agents, draft content, summarize information, analyze data, classify tasks, create reports, generate deliverables, hold conversations with your customers in writing and by voice, and automate approved workflows.
AI Office may include Linda orchestration, AI agents, third-party AI models, tool integrations, connected channels and customer-configured workflows.
2. AI Office supports, but does not replace, human judgment
AI-generated outputs may be wrong, incomplete, biased, outdated or unsuitable for your business, legal, financial, employment, medical, security or compliance needs.
You are responsible for reviewing AI outputs before relying on them, sharing them, sending them to customers or contacts, updating business systems, making decisions or taking external action.
Do not use AI Office as the only basis for decisions affecting legal rights, employment, credit, insurance, healthcare, safety, security, finance or access to services unless your organization has approved that use and implemented legally required human review.
3. Agent capability labels matter
Agent cards and workflow screens show capability labels: Live Specialist, Generic Advisory, Needs Connector, Beta, Coming Soon and Disabled.
A Generic Advisory agent can draft, summarize, reason or recommend, but is not a fully specialized operational agent. A Needs Connector agent cannot access or update the relevant system until your organization authorizes and configures the connector. Beta and Coming Soon features may be incomplete or experimental. Coming Soon and Disabled agents cannot be assigned to paid work.
4. What data may be processed by AI
Depending on what your organization enables, AI Office may process task briefs, prompts, messages, documents, files, records, leads, invoices, support tickets, notes, Work Runs, artifacts, comments, approvals, usage data, connector data, and the content of conversations held through connected channels, including written transcripts of telephone calls.
Notes you add to train an agent are stored as text and retrieved by keyword relevance and recency.
Only provide data you are authorized to use. Do not submit sensitive, confidential, regulated, personal, client, employee, health, biometric, financial or children's data unless your organization has approved that use and the correct safeguards are in place.
5. Agents talk to real people
Where your organization enables a channel, agents may hold conversations with your contacts and customers on your behalf:
- on your website, through the chat widget you embed;
- on WhatsApp, where you connect a business number;
- by email, where you route an address to AI Office;
- by telephone, where you connect a phone number. The AI answers the call, speaks aloud, and a written transcript is stored against the resulting ticket.
Your organization is responsible for consent, opt-outs, call recording and transcription notices, and compliance with messaging, telecommunications and consumer protection law in each market you operate in. Where required, tell people at the start of the conversation that they are speaking with an automated system. The product provides configurable greetings for exactly this purpose.
Review agent-drafted messages before they are sent wherever approval controls are available.
6. Third-party AI and service providers
AI Office uses third-party providers to process your requests, currently including Anthropic and, where you select one of their models, OpenAI for AI model processing; ElevenLabs for speech; Twilio for telephony; Meta's WhatsApp Business Platform for WhatsApp; Resend for email; and Stripe for payments. This involves sending prompts, context, files, metadata, outputs, usage information and, on the voice channel, speech and transcripts to those providers to complete the requested task.
Customer Content is not used to train public foundation models. The full subprocessor list, including hosting and infrastructure providers, is at [subprocessor page URL].
7. Work Runs, logs, costs and admin visibility
AI Office records Work Runs, agent steps, artifacts, approvals, comments, source references, model usage, token usage, tool usage, channel messages, call transcripts, costs and audit events for transparency, debugging, billing, security and governance.
Workspace administrators may view your AI activity, prompts, outputs, Work Runs, deliverables, channel conversations and related audit records, depending on your organization's settings.
8. Responsible use
Do not use AI Office to create malware, phishing, spam, scams, impersonation, deception, harassment, discrimination, unlawful profiling, unauthorized surveillance, illegal content or content that violates third-party rights. Do not send unsolicited messages or call contacts who have opted out.
Do not use AI Office to bypass security controls, extract secrets, scrape data unlawfully, make unauthorized decisions about people, or process data you are not allowed to process.
AI-generated content should be labelled or disclosed where required by law, policy, contract or professional standards.
9. Reporting an AI issue
Report inaccurate, unsafe, biased, harmful, unauthorized or suspicious AI outputs to your workspace administrator, or contact security@w3turn.com or privacy@w3turn.com.
Where available, use the product's feedback controls: request changes, reject an approval, or agent feedback.
Suggested in-product short notice
W3TURN AI Office uses AI agents to generate plans, drafts, reports and recommendations, and can talk to your customers by chat, WhatsApp, email and telephone through channels you enable. Outputs may be inaccurate and must be reviewed before use. Calls are transcribed. Do not submit sensitive or unauthorized data. Your organization may log AI activity for security, billing and governance. [Learn more]