Press

Press resources for Keel

Permits have made consequential human work safer for decades. As AI agents gain access to systems, data, budgets, payments, and workflows, organizations need the same discipline: define authority before actions occur, then prove what happened afterward.

About Keel

The control plane for accountable AI action.

Keel helps companies create a safer operating environment for AI agents. A governed action gets a permit before it runs, its result is attached afterward, and the record can be verified independently.

Permits before action

Permits have governed consequential human work for decades. Agents need the same explicit boundary before they touch systems.

Safer agent environments

Give teams a controlled space to deploy agents: clear rules, scoped authority, and room to improve without silent risk.

Accountability that improves

Fail closed when an action exceeds scope, then use the record to correct behavior and raise the bar.

Proof after the fact

Bind each governed action to durable evidence an auditor, customer, or operator can verify independently.

Company description

Keel is an AI authorization and accountability platform. It helps organizations decide what AI agents are allowed to do before they act and create independently verifiable evidence afterward.

What Makes Keel Different

Other systems build, route, run, scan, observe, or report on AI. Keel defines execution authority before action and records evidence that can be checked afterward.

Most AI tools	Keel
Agent builders and orchestrators	Controls what agents may do
Gateways and proxies	Authorizes execution before dispatch
Observability and monitoring	Blocks, challenges, and proves decisions
Evals and quality scoring	Uses policy context to govern execution
Budget alerts and FinOps dashboards	Denies or throttles before provider spend
Content filters and security scanners	Enforces authority at the action boundary
Identity and IAM	Applies identity context to execution authority
Policy engines and approval tools	Turns policy decisions into permit-bound action records
RAG, memory, and MCP registries	Governs retrieval and tool-call authority
GRC and compliance workflows	Supplies tamper-evident AI execution evidence

Agent builders and orchestrators

Controls what agents may do

Gateways and proxies

Authorizes execution before dispatch

Observability and monitoring

Blocks, challenges, and proves decisions

Evals and quality scoring

Uses policy context to govern execution

Budget alerts and FinOps dashboards

Denies or throttles before provider spend

Content filters and security scanners

Enforces authority at the action boundary

Identity and IAM

Applies identity context to execution authority

Policy engines and approval tools

Turns policy decisions into permit-bound action records

RAG, memory, and MCP registries

Governs retrieval and tool-call authority

GRC and compliance workflows

Supplies tamper-evident AI execution evidence

Christian Munoz

Founder

San Francisco, CA

Christian Munoz is the founder of Keel. Born and raised in San Francisco, he spent more than a decade as an educator and communication coach, working with business leaders, engineers, and professionals across the United States and Ukraine. His experience designing environments where people can learn, improve, and operate within clear boundaries influenced Keel's approach to AI governance: defining authority before action and creating verifiable evidence afterward.