You, your AI agent and Avokata

Three roles

• You, the operator: set the task, explain the facts and relevant date, and decide how to use the result. Review important conclusions before acting on them.
• Your AI agent: plans the work, chooses whether and how to call tools, reads the responses and drafts the answer. Its abilities depend on the model, app, available tools, instructions and conversation context. Memory, skills and your guidance can also affect its work.
• Avokata: supplies the research tools and the material they return. We maintain our tools, datasets and service, disclose known coverage and freshness limits, and investigate and correct faults within Avokata.

These are roles in a working process, not a transfer of our legal obligations to you.

Why use Avokata with an agent?

Your agent can retrieve legal material, follow references and check the wording and date against available sources, rather than relying only on its model’s memory. That gives you material you can inspect and cite. The benefit depends on the agent actually using suitable tools and on what Avokata holds.

Avokata is not an official register, a law firm or a guarantee of a correct final answer. Our search, extraction, relationships and source material can contain errors or gaps. Available tools do not guarantee that your assistant will use them, use every relevant source or interpret the result correctly.

Check before relying on a result

AI can make mistakes, including inventing a reference or overlooking relevant material. Before a consequential decision, verify the important claims, citations and wording against authoritative sources, the applicable jurisdiction and date, and your actual facts. Check the relevant edition, later changes and whether a decision is final where that matters. Obtain appropriate professional advice when needed.

An empty search result does not mean that no relevant law or decision exists. A missing document, incomplete history, unavailable tool or stale source is a limitation to resolve, not a legal conclusion. Ask the agent to identify what it checked, what remains uncertain and what it could not obtain.

What we can fix

We are responsible for maintaining and improving what is within Avokata: tool behaviour, dataset quality and coverage, disclosed source dates, performance and availability. We do not promise complete coverage, error-free data or uninterrupted access. Disclosing a limitation does not excuse a fault in our own service.

Your assistant and the app you use are provided separately. We do not control their reasoning, tool selection, instructions, memory or handling of your conversation. A poor result may involve the agent, Avokata, the source or several of them; we investigate evidence rather than assuming it is the user’s fault.

Tell us what went wrong

Use Avokata’s feedback tool or write to hello@avokata.com. Give the tool name, public source identifier, relevant date, and what you expected. Share only what is needed; omit client details, secrets and other sensitive information.

This guide forms part of the Terms of Use. It does not exclude rights or responsibilities that applicable law does not allow us to exclude. The privacy notice explains our handling of account data; your AI provider has its own terms and privacy practices.