Navital Gym · answers
How do you evaluate an AI agent for Indian finance and the RBI FREE-AI framework?
RBI's FREE-AI report sets out principles and recommendations for responsible AI in the financial sector, including a pillar on assurance. A test like Navital Gym does not certify compliance with it. What it gives you is a record of how an agent behaved on specific UPI and GST tasks, graded by written rules, including whether it attempted actions it was not allowed to take.
As of 12 November 2026 · Navital Gym v1.1.0
What a test can show
- Whether the agent attempted a forbidden action, counted even if the simulated bank refused it.
- Whether it refused an impossible request and gave the real reason.
- Which rule failed, by id, so a reviewer can read the failure.
- The same result every run: the grading has no AI judge.
Written rules, no AI judge
- Written rules recompute the answer from the sandbox. Nothing the agent claims is trusted, and no AI model is the judge.
- A task counts as solved only when every required rule passes, on one attempt.
- Many tasks have no valid answer. The agent has to say so and give the real reason.
Sources
FREE-AI is the Reserve Bank of India's committee report on the Framework for Responsible and Ethical Enablement of Artificial Intelligence (published 13 August 2025). NPCI has also published a banking-agent benchmark.
Have us test your agent
- Private evaluation, $750 per agent version: your agent on our private held-out tasks, with a written report within 5 working days after setup.
- Pilot, from $2,500: 20-30 tasks built around one flow of yours, with grader, reference solutions and a short report.
- Training Pilot, from $7,500: we train an open model you choose on a world and test it on held-out tasks.
Prices in USD, plus applicable taxes (none for clients outside India). Scope is written down before work starts.
Questions
What is RBI FREE-AI?
The report of the Reserve Bank of India's committee on a Framework for Responsible and Ethical Enablement of Artificial Intelligence in the financial sector, published on 13 August 2025. It sets out principles and recommendations under six pillars, one of which is assurance.
Does Navital Gym certify FREE-AI compliance?
No. Navital Gym is a test of agent behaviour on simplified, synthetic tasks. It makes no compliance or certification claim and is not legal advice.
What can a Navital Gym result show for a finance agent?
How the agent behaved on specific tasks: whether it tried a forbidden action, refused an impossible request, and which written rule failed. It is evidence about those tasks, not about your whole system.
Can you test my finance agent privately?
Yes. A private evaluation runs your agent on held-out tasks for $750 per agent version. We sign your NDA.
What this does not show
- No compliance, certification or regulatory-alignment claim is made. Tasks are synthetic and the rules are simplified.
- The tasks cover UPI and GST workflows, not lending, KYC at a bank or every FREE-AI area.
- Not affiliated with RBI, NPCI, GSTN or any bank.
Navital GymLeaderboardHow to test an AI agent on UPI and GST workflowsGST agent benchmark: Navital Gym GST DeskHow to evaluate Hindi and Hinglish customer support agents