Kamiwaza 1.0: secures AI orchestration for regulated industries

Cross-functional connections

Enterprise and government teams regularly work in cross-functional groups where sensitive data must be shared within a project, but not universally. 

Usually, that means either restricting access so broadly that collaboration suffers, or opening it up in ways that create security and compliance risk – neither is acceptable in regulated environments.

The release introduces three new major features:

• Kamiwaza Workrooms, a new governed collaboration environment where enterprise teams and their AI agents work within their own access boundaries.
• Infrastructure built on Chainguard’s hardened, zero-to-low CVE container images, designed for security and compliance.

• Kaizen, the platform’s AI agent, upgraded with expanded capabilities for multi-modal analysis and output.

Luke Norris, CEO and co-founder of Kamiwaza says that together, these features enable enterprise teams to collaborate with AI across sensitive data in a way that was not previously possible, with strict boundaries enforced at the platform level.

“Regulated industries have been clear about what they need from AI: keep our data where it is, respect our security boundaries, and give us full visibility into what the AI is doing,” said Norris. “Kamiwaza 1.0 builds on the distributed data and security foundation our enterprise and government customers already rely on with hardened infrastructure, governed team collaboration, and a more capable agent that works across all of it.”

Kamiwaza Workrooms creates policy-bound spaces where team members and their AI agents operate within their individual access rights.

Each Workroom contains its own data and tools, accessible only to those with the appropriate permissions. The platform enforces these boundaries at the architecture level, not through manual policy exceptions or agent-level filters.

Every action is fully auditable.

Security-hardened infrastructure 

According to Norris and team, most enterprise AI platforms are built on standard open source container images that accumulate vulnerabilities over time. Security teams are left to identify, triage, and patch those vulnerabilities in infrastructure they didn’t build and don’t control. This becomes an unsustainable model as AI workloads move into production.

Kamiwaza 1.0 addresses this by using Chainguard Containers, hardened container images purpose-built for security and compliance. Unlike standard open source images, Chainguard’s container images are minimal by design and rebuilt continuously, delivering zero known vulnerabilities, high-quality SBOMs, and verifiable signatures. FIPS-ready versions are also available for federal deployments. 

“Workroom, Chainguard and Kaizen each solve a distinct problem, but they add up to something bigger,” said Matt Wallace, CTO and co-founder of Kamiwaza. “Teams can collaborate with AI across sensitive data without anyone – human or agent – seeing more than they should. The infrastructure they’re running on has zero known vulnerabilities from day one. And the agent connecting it all understands the full context of their data as opposed to its fragments. That combination doesn’t exist anywhere else.”

Kaizen, the platform’s flagship AI agent, connects to internal data sources across an organisation’s systems through the Kamiwaza Context Manager, so its outputs are informed by the full data landscape rather than a single silo. A new skills library lets enterprise teams define which capabilities are available to the agent and under what conditions.

Kamiwaza 1.0 is generally available as of today.