Responsible AI at Tagile

Tagile's background agents gather information from public sources and the systems a customer chooses to connect. AI is used to filter out noise, identify changes that are relevant to the customer's relationships, summarize what happened, and draft a possible next step, for example a suggested email or comment.

• Monitors publicly available information about the companies and people you choose to follow.
• Filters, ranks and summarizes changes so only relevant signals reach you.
• Prepares draft suggestions you can edit, discard or approve.

• It does not send messages, post comments or make connection requests on its own.
• It does not change data in your external systems on its own.
• It does not make binding business decisions for you.

Tagile is a decision support system. A person always has the last word: nothing is sent, published or updated in your systems without a human reviewing and approving it first.

All AI generated suggestions are drafts. Users can edit, discard or approve them, and are expected to review content before it is used or shared externally.

Generative AI can sometimes produce incorrect or misleading information. We work to reduce this by grounding output in gathered information and, where possible, showing the sources behind a summary, but users remain responsible for verifying critical information before acting on it.

Where possible, Tagile shows the sources a summary is based on, so users can check the underlying information themselves.

Recommendations are suggestions, not instructions. They reflect the information available to the system at the time and should be weighed against your own knowledge of the relationship and situation.

Tagile does not train public or general purpose AI models on your Customer Data or your interactions with the platform. Your business knowledge stays yours.

Tagile uses established external AI providers to process text for analysis and summarisation, accessed through their business APIs. Our aim, reflected in our provider selection and settings, is that data sent for processing is not used to train providers' public models. Providers are listed on the Subprocessors page.

We aim to send external AI providers only the information needed for the specific analysis, and to avoid including unnecessary personal data.

• Human oversight: people approve before anything leaves the platform.
• Purpose limitation: AI is used to support B2B relationship work, nothing else.
• Transparency: we are open about what the AI does and its limitations.
• Regulatory awareness: Tagile assesses the requirements that apply to its use of AI and reviews its practices as the EU AI Act and related guidance develop.

Customers must not use Tagile for automated mass outreach, must review AI generated content before use, and must comply with the acceptable use rules in the Terms of Service.

Questions about our use of AI? Contact hello@tagile.ai.