Idle Work Usage Policy

Effective: September 22, 2026 · Last updated: September 2026

This Usage Policy sets out what the Idle Work platform and AI employee agents may and may not be used for. It is incorporated by reference into the Idle Work Commercial Terms of Service and applies to the Customer and to every User the Customer allows onto its agents. An agent acts on the Customer's instructions, so the Customer is responsible for what it is asked to do.

1. Permitted Use

The Services are for business use by the Customer, to perform business functions for the Customer's own organisation and its own customers and contacts. The Customer may build the agent into products and services it offers to its own Users, subject to these terms. The Services are not offered to consumers or for personal use.

2. Prohibited Uses

The Customer and its Users must not use the Services, or allow them to be used, to:

• Do anything that is a criminal offence, gives rise to civil liability, or violates any applicable law or regulation, including data protection, consumer protection, employment, financial services and export laws.
• Send spam, chain messages, or any unsolicited bulk communication by any channel, including chat messages and email.
• Message any person who has not consented to be contacted, has opted out, or may not lawfully be contacted for that purpose.
• Communicate with a minor, or knowingly collect information about a minor.
• Impersonate any person or organisation, or misrepresent who is speaking, who the Customer is, or what the Customer is affiliated with.
• Deceive a recipient about whether they are dealing with an automated system where a law, a platform rule, or the recipient's direct question requires disclosure.
• Share, send or generate content that is unlawful, abusive, threatening, harassing, defamatory, obscene, discriminatory, or that infringes anyone's intellectual property or privacy.
• Distribute malware or any harmful, disruptive or destructive code or file.
• Disrupt, overload, probe, or interfere with the Services, the machines and networks behind them, or other Customers' use of them.
• Attempt to gain unauthorised access to the Services, to another Customer's agent or data, or to any system the agent is connected to beyond the access the Customer has been granted.
• Circumvent, disable or tamper with any approval gate, permission scope, rate limit, channel restriction or safety control in the Services.
• Extract, scrape, copy, resell, sublicense or redistribute the Services, the agent, its skills, its prompts or the data Idle Work provides, except as the Services expressly allow.
• Use the Services to build, train or benchmark a competing product or model, or help anyone else do so.
• Reverse engineer, decompile or otherwise attempt to derive the source or workings of the Services.

3. Messaging Rules

When the agent sends messages, the Customer is the sender of record. The Customer must:

• Obtain, and keep evidence of, each recipient's consent to be contacted by that channel, as the Telephone Consumer Protection Act, CAN-SPAM, and comparable laws and industry codes require.
• Honour opt-outs promptly, including a reply of STOP, an unsubscribe link, or a plain request to stop, and not contact that person again on that channel.
• Follow the rules of the channel provider: Meta's WhatsApp Business and Messaging policies, Telegram's terms, Discord's terms and community guidelines, Slack's and Microsoft's acceptable use policies, and the Customer's email provider's sending policies. A channel provider may restrict or suspend a number or account that breaks its rules, and that is a matter between the Customer and the provider.
• Send only content the Customer would be willing to have attributed to it, since it will be.
• Use SMS only where Idle Work has enabled it for the Customer in writing; SMS is a pilot channel and is not part of the standard Services.

4. Human Oversight

Outputs can be wrong. The Customer is responsible for deciding where human review is required before an Output is used, sent or acted on, and for telling its Users that factual statements in Outputs must be checked. In particular:

• Sending, posting and writing to connected systems are unlocked individually in each agent's settings. A new agent starts by watching, drafting and reporting. The Customer decides what it may do on its own and what comes back for a yes.
• Purchases made through Link are approved by the Customer in the Link app before any payment credential is issued. The Customer is responsible for every purchase it approves.
• Price changes, invoices, campaigns and social posts are shown to the Customer before they go out unless the Customer has expressly turned that gate off for that action.
• Outputs are not professional advice. Document reviews, scope reviews, findings, financial summaries and similar Outputs are aids to the Customer's own judgement and are not legal, engineering, accounting, tax or medical advice. Decisions that require a licensed professional still require one.

5. High-Risk Uses

The Customer must not use the Services as the sole basis for any decision that has a legal or similarly significant effect on a person, including decisions about employment, housing, credit, insurance, education, or access to essential services, and must not use the Services for medical diagnosis or treatment, for anything involving weapons or controlled substances, or for any activity that presents a risk to life or safety. Where a use in a regulated industry needs additional controls, the Customer must agree those with Idle Work in writing before starting.

6. Accounts and Credentials

• The Customer is responsible for all activity under its account and on every agent it runs, including by its Users.
• Account credentials must not be shared. Each person who needs access gets their own login.
• Connected-account credentials (Google, Microsoft 365, Dropbox, Meta, Plaid, point-of-sale and email-campaign systems) are granted through each provider's own authorisation flow. The Customer must have the authority to grant them and must revoke any it no longer wants the agent to hold.
• The Customer must tell Idle Work promptly if it believes its account or agent has been compromised.

7. Idle Work's Rights

Idle Work may ask the Customer for reasonable information to verify its identity and its use of the Services, and the Customer must cooperate. Idle Work may suspend all or part of the Services, as the Commercial Terms describe, where it reasonably believes this policy is being breached, where there is a risk to the Services or to other Customers, or where a channel provider or subprocessor suspends the underlying service. Idle Work will use reasonable efforts to give notice, and will restore access as soon as the cause is cured, where it is curable. Idle Work does not monitor the content of Customer conversations as a matter of course, and this policy does not oblige it to.

8. Reporting Abuse

If you believe an Idle Work agent is being used in breach of this policy, or you have received a message from one that you did not consent to, contact us at the address below with the number or account that contacted you and a copy of the message. We investigate every report.

9. Changes to This Policy

We may update this Usage Policy. Changes take effect 30 days after they are posted here, except changes required by law or by a channel provider's rules, which take effect on posting.

10. Contact