Responsible intelligence.

This AI Policy explains the principles and practices that govern artificial intelligence across the Taskclan platform, including Taskclan Intelligence, Cloud, Studio and Engine, the T1 model family, and every product built on the platform.

It applies to anyone who uses AI features through Taskclan, and works alongside our Terms of Service, Acceptable Use Policy, Data Processing & Protection Policy and Privacy Policy.

We build AI to be useful, governable and trustworthy. Five principles guide every decision:

• Human oversight. AI proposes and executes; people stay accountable. High-consequence actions are designed to keep a person in the loop.
• Governance by architecture. Identity, policy, permissions and scope are enforced by the system, not by good intentions.
• Transparency. We are clear about when you are interacting with AI, what it can do, and its limits.
• Model independence. We route to the best available intelligence rather than locking you into a single provider.
• Continuous evaluation. Quality, safety, rights and policy are checked before outputs ship, and outcomes feed back into the system.

T1 profiles

Taskclan exposes intelligence through T1 profiles, Core (fast, efficient), Flow (adaptive orchestration) and Max (maximum reasoning depth). A profile is not a single model; it is a routing strategy that selects the right model, tools and reasoning depth for each task.

Model routing

Behind T1, requests may be served by leading open and proprietary frontier models, which we evaluate continuously for quality, latency, cost and safety. The specific model used for a given request may change as the frontier moves.

Third-party providers

Some intelligence is provided by third-party model providers acting as our subprocessors. See our Subprocessors list for who they are and where they operate.

We do not use your prompts, inputs or generated outputs to train foundation models, and we do not sell them.

• Your content stays yours. Content you submit is processed to deliver the feature you requested and is governed by our Data Processing & Protection Policy.
• No silent training. We will not use identifiable customer content to train or fine-tune models without a separate, explicit opt-in.
• Provider terms. Where a request is served by a third-party model, that provider processes it under contractual terms that restrict training on your data.

AI systems are probabilistic. They can be confidently wrong, produce out-of-date information, or generate content that appears plausible but is inaccurate ("hallucinations").

• Verify before you rely. Outputs are a starting point, not professional, legal, medical or financial advice.
• Context matters. Quality depends on the inputs, instructions and grounding data you provide.
• Not a person. AI features do not have beliefs, intentions or judgment, even when they sound like they do.

Taskclan Engine includes an evaluation layer that checks output quality, policy compliance, rights and outcome before anything ships. For consequential or irreversible actions, workflows are designed to require human confirmation.

Enterprise customers can configure approval paths, permissions and regional rules so that AI operates within the boundaries their organization sets.

You may not use Taskclan AI for activities that are illegal, harmful or abusive, including generating disinformation, non-consensual or exploitative imagery, malware, or content that infringes others’ rights. A full list is in our Acceptable Use Policy, which is incorporated into this policy by reference.

As between you and Taskclan, and subject to our Terms of Service, you own the content you create with our tools. You are responsible for ensuring your inputs and your use of outputs comply with applicable law and third-party rights.

Because generative models can produce similar outputs for different users, we do not warrant that any output is unique or free of third-party claims. Review outputs before commercial use.

If you encounter AI output or behavior that seems unsafe, biased, infringing or otherwise concerning, contact us at hello@taskclan.com or through our contact page. We investigate reports and use them to improve the system.