Solcis releases tools that check each AI agent task against rules

On 21 October 2026, Solcis, a spin-out of the firm 6clicks, released Solcis Compliance Studio and the Solcis Policy Engine. They turn company policies into rules for AI agents. They check each task of an agent against the rules when it happens, and they keep a record for auditors. The two tools are available at this time.

How it works

Compliance teams write and manage policies in Studio. The Policy Engine checks each task in two methods. Fixed rules give the same result each time. A small language model (a small AI model) judges unclear conditions. Solcis also works with NVIDIA OpenShell and the Microsoft Agent Governance Toolkit.

The proof

The one number is the speed claim of less than 200 milliseconds. Solcis gave no test results, no customer names and no data on how correctly the AI checks work.

What is not known

Solcis software runs only on the systems of the customer. The code of the Engine is open to audit, with the Functional Source License. Studio is sold by subscription or perpetual licence. The price is not known. The results with agents in use are not known.

Insight: Rule engines are a crowded layer and the record is the prize

Solcis connects to engines from NVIDIA and Microsoft. In the week before, AWS released Strands Box, a sandbox with a policy language, and Strands Decider 2B, a small decision model. Many open engines apply rules at this time. The main value seems to be with the firm that holds the policy and the audit record. It seems a race to hold that record. Find who holds yours.