AI Usage Policy

Konversation.io uses artificial intelligence, including large language models provided by third-party subprocessors such as OpenAI and Anthropic, to help businesses automate and improve customer conversations. This Policy explains the limitations of these features, the responsibilities of customers, and the safeguards we apply.

AI features are powerful but imperfect. They generate output based on patterns in data and do not possess judgment, understanding, or awareness. AI Output may be incomplete, outdated, biased, or factually incorrect, may misinterpret context, and may not be appropriate for every situation. AI features should be treated as a tool to assist humans, not as a substitute for professional judgment.

You must ensure appropriate human oversight of AI features, particularly where output could have a significant effect on individuals. You should implement human review before AI Output is used for decisions involving:

• legal, financial, medical, or safety-related matters;
• eligibility for products, services, credit, or employment;
• any decision producing legal or similarly significant effects on an individual; and
• sensitive communications where accuracy is critical.

You are responsible for:

• configuring, testing, and monitoring your AI agents and automations;
• providing accurate knowledge-base and training content;
• disclosing to your end users, where appropriate, that they are interacting with an AI system;
• ensuring your use complies with applicable AI, consumer-protection, and data-protection laws; and
• obtaining any consents required to process inputs through AI features.

You must not use AI features to:

• generate unlawful, harmful, hateful, harassing, or deceptive content;
• create disinformation, deepfakes, or impersonations intended to mislead;
• produce sexual content involving minors or non-consensual content;
• provide unqualified professional advice presented as authoritative;
• make fully automated decisions with significant effects without lawful basis and safeguards; or
• attempt to bypass, manipulate, or disable AI safety controls.

We expect customers to use AI responsibly and ethically, with transparency toward end users, respect for privacy and human dignity, and fairness in outcomes. You should avoid uses that could deceive, manipulate, or unfairly disadvantage individuals, and should consider the impact of automation on the people it affects.

Inputs and outputs of AI features may be processed by our AI subprocessors solely to deliver the requested functionality. We do not use customer conversation content to train our own general-purpose models without authorisation, and we contractually require our AI subprocessors not to train their foundation models on data submitted through our integrations except as permitted. Aggregated and de-identified data may be used to monitor and improve quality and safety. See our Privacy Policy and Subprocessors page.

We implement safeguards designed to reduce harmful or abusive AI use, which may include content filtering, rate limiting, abuse detection, prompt and output monitoring, and reliance on the safety systems of our model providers. These measures are not infallible, and we continue to refine them as technology and best practices evolve.

We may restrict, suspend, or terminate access to AI features, in whole or in part, if we reasonably believe they are being used in violation of this Policy, our Acceptable Use Policy, applicable law, or the terms of our AI providers, or where necessary to protect the safety, security, or integrity of the Services. This Policy was last updated on June 15, 2026. For questions, contact team@konversation.io.