autoDM / Legal

AI Supplementary Terms

Effective 23 October 2026 · PT Kalachakra Wisnu Indonesia (“autoDM”)

In shortOn AI-enabled plans, autoDM can draft and send replies to Instagram DMs using a large language model provided by Anthropic. The AI answers only from information you give it, your data is never used to train models, and you stay responsible for what is sent in your name. AI output can be wrong, so keep a human in the loop and make sure the hand-off to your team works.

1. When these terms apply

These AI Supplementary Terms (“AI Terms”) apply whenever you enable or use any AI-enabled feature of autoDM (“AI Features”). They supplement the Terms of Service and the Acceptable Use Policy; if there is a conflict about AI Features, these AI Terms prevail. Capitalised terms have the meaning given in the Terms of Service.

AI Features are available only on AI-enabled Plans. Plans without AI Features use rule-based automation only (keyword replies, templates and quick-reply menus) and are not subject to these AI Terms.

2. What the AI Features do

Depending on your Plan and configuration, AI Features may:

  • Generate replies to incoming Instagram direct messages, including follow-up messages in conversations that began with a comment-to-DM or story-reply trigger;
  • Detect intent in incoming messages so that the right Automation, answer or team member is chosen;
  • Hand off to a human when the AI cannot answer confidently, when a Subscriber asks for a person, or when a rule you set is triggered.

The AI generates answers only from the Knowledge you provide in the Service (your FAQs, product catalogue, business details, tone instructions and similar content), together with the current conversation. It is not connected to the open internet and is not designed to answer questions outside your Knowledge.

3. Our AI provider

AI Features are powered by large language models provided by Anthropic, PBC (“Anthropic”) through its commercial API. To generate a reply we send Anthropic the relevant parts of your Knowledge, the recent messages in the conversation, and system instructions we write. Anthropic processes this data under a commercial agreement with us that:

  • prohibits Anthropic from using your Knowledge, Subscriber messages or generated output to train or improve its models;
  • limits Anthropic to processing the data in order to return a response to us; and
  • requires appropriate security and confidentiality measures.

Anthropic may process data outside Indonesia. We treat Anthropic as a sub-processor under the Data Processing Addendum. We may change AI provider or model; if we do, the replacement will be bound by equivalent commitments and we will update our Service Providers page.

4. Your Knowledge and Inputs

“Inputs” are the Knowledge you provide and the Subscriber messages the AI reads. “Outputs” are the replies, classifications and other content the AI produces. As between you and us, you own your Inputs and, to the extent permitted by law, the Outputs generated for your Account.

You are responsible for your Inputs. You confirm that:

  • your Knowledge is accurate, lawful and up to date, and you will keep it so;
  • you have the rights needed to use your Knowledge in this way;
  • your Knowledge does not include special categories of personal data (such as health, religion or biometric data) or data about children unless you have a lawful basis and have told us in writing; and
  • you will not include instructions in your Knowledge that would cause the AI to breach these AI Terms or the Acceptable Use Policy.

5. Nature of AI output

Large language models are probabilistic. Even with accurate Knowledge, Outputs may be incomplete, out of date, misleading, inappropriate, or simply wrong, and may vary between identical requests. Outputs may not be unique to you; other Customers may receive similar Outputs. Accordingly:

  • Outputs are not professional advice of any kind and should not be relied on as such;
  • you must review your configuration and sample conversations before going live, and monitor conversations after launch;
  • you must not present Outputs to Subscribers as statements of fact about prices, availability, legal rights, health, safety or finances without a mechanism for human verification;
  • you are responsible for correcting any error the AI makes with a Subscriber, including honouring a price or promise the AI made if applicable consumer law so requires; and
  • we do not warrant that Outputs will be accurate, appropriate or fit for any purpose.

6. Human oversight and hand-off

You must keep a working route for Subscribers to reach a human. In practice this means you must:

  • configure the hand-off feature so that unanswered or escalated conversations reach a member of your team;
  • respond to handed-off conversations within a reasonable time; and
  • allow Subscribers to request a human at any point, and never configure the AI to refuse or discourage such a request.

7. Transparency to Subscribers

You must not use AI Features to deceive Subscribers about whether they are interacting with an automated system. At a minimum, the AI must not claim to be a human when a Subscriber asks, and you must disclose the use of automation wherever law, Meta’s policies or your own industry rules require it. We recommend a short notice at the start of AI-handled conversations.

8. Prohibited uses of AI Features

In addition to the Acceptable Use Policy, you must not use AI Features to:

  • generate content that is illegal, harmful, hateful, sexually explicit, or otherwise prohibited by the Acceptable Use Policy;
  • provide medical, legal, financial, insurance or safety-critical advice without human review;
  • make or communicate decisions that have legal or similarly significant effects on a Subscriber (such as eligibility for credit, employment, housing or insurance);
  • profile, score or manipulate Subscribers, or infer special categories of personal data about them;
  • attempt to extract the system instructions, prompts or model weights, or to jailbreak, probe or attack the model;
  • use Outputs to develop or train a competing model or service; or
  • breach Anthropic’s Usage Policy as published at anthropic.com/legal/aup, to the extent it applies to end uses of its models.

We may apply automated safety filters to Inputs and Outputs. These may occasionally block legitimate messages; if that happens, contact support.

9. Usage limits

AI-enabled Plans include a monthly allowance of AI-handled conversations or messages as described on our pricing page. If you exceed your allowance we may pause AI replies (your rule-based Automations continue), offer an upgrade, or apply fair-use throttling. We may also throttle usage that is abnormally high or that we reasonably believe is automated or abusive.

10. Data handling

Inputs and Outputs are processed as Customer Content and Subscriber Data under our Privacy Policy and Data Processing Addendum. We never use your Inputs or Outputs to train AI models, and we do not permit our providers to do so. We retain conversation logs for the periods described in the Privacy Policy so you can review AI performance; you can delete them earlier from the Service.

11. Intellectual property

You may use Outputs for any lawful purpose connected with your business. Because Outputs are machine-generated, their protection under copyright law is uncertain and we make no promise that you can enforce rights in them. Outputs may inadvertently resemble third-party content; you are responsible for checking Outputs you publish outside the Service.

12. Liability

The limitations and exclusions of liability in the Terms of Service apply to AI Features. In addition, and to the fullest extent permitted by law, we are not liable for any loss arising from reliance on an Output, from an Output that is inaccurate or inappropriate, or from the unavailability of the AI provider’s service.

13. Changes

AI technology and its regulation are developing quickly. We may update these AI Terms, change models or providers, or change how AI Features work, and will give notice of material changes under the Terms of Service. We may also suspend or withdraw an AI Feature at short notice if required by our provider, by law, or to protect Subscribers.

14. Contact

Questions about AI Features: support@autodm.kalachakra.io. Privacy questions: privacy@autodm.kalachakra.io.