Businesses run by agents. Actions you can prove.
Action State Group is an AI product studio. We build agent-operated companies for businesses of all sizes AND the open trust infrastructure the agent economy needs when AI agents start taking real-world actions.
Meet GetGopher →
Filed as an individual IETF Internet-Draft · Member, Agentic AI Foundation · Open verifier, Apache-2.0 · Built with the IETF SCITT and Linux Foundation communities.
What is Action State Group
Action State Group (a.k.a. Action State) is a studio for AI-native companies. We design, build, and operate ventures where agents do the work — beginning with professionals in high-trust fields and evolving to businesses of all sizes.
Operating businesses on agents taught us the industry's next problem firsthand: when agents send the email, pay the invoice, or file the form, a log entry isn't proof. Disputes, audits, and insurance all eventually ask the same question — what did the agent actually do? So alongside our products, we build open, vendor-neutral infrastructure for agent actions, and we contribute that work to the standards community.

Studio approach
Single platform that builds and operates multiple personas, verticals, and workflows.
Agentic systems
Using agents as the new form of computing…from state to action.
High-trust work
Designed for businesses where consequential actions are front and center to their operations.
My Perspective
For most of my career, I have worked in open source, distributed systems, and cloud infrastructure; tech that gave developers enormous leverage.
Agentic systems are the next step in that story: the same kind of leverage, extended to people who operate on judgment and critical decisions. But leverage that acts in the world needs something developer tools never did — accountability. The lesson I keep relearning from open source (Presto Foundation then; the agentic ecosystem now) is that infrastructure everyone must trust can't be owned by one vendor. That's why our trust tooling is open, vendor-neutral, and headed for community governance — and why we build it in the open, with the IETF and Linux Foundation communities.
What we're building
Agent Action Capsule — proof of what an agent did
An open IETF Internet-Draft and reference implementation for a tamper-evident, content-private, independently verifiable record of a single agent action — checkable by anyone, without trusting the operator. Built on the SCITT/COSE transparency standards, with an open verifier you can run in under a minute.
GetGopher
A managed AI operator for business owners — delivered over text message, run by us, governed by deterministic checks with a human approval on anything consequential. Built for people who run their business from their phone.
scitt-cose
The neutral substrate beneath the capsule — an Apache-2.0 standalone verifier for anyone's SCITT signed statements and receipts (treats the payload as opaque bytes), plus a free, stateless hosted verifier at verify.actionstate.ai.
agent-action-capsule
The reference implementation of the Agent Action Capsule spec — a tamper-evident, content-private, independently verifiable record of a single agent action. Built on SCITT/COSE transparency standards.
capsule-emit
A library for emitting Agent Action Capsules from your agent code — structured, signed, and ready to verify.
Field Notes on the Agent Economy
Working notes from operating businesses on agents and building the trust layer they need — what works, what breaks, and what the ecosystem is missing.
Logs prove nothing — and the web already fixed this once
Field Notes on the Agent Economy · Action State · June 2026
This week we open-sourced scitt-cose, a standalone verifier for IETF SCITT receipts and signed statements, along with a free hosted verifier at verify.actionstate.ai. This post is the story of why a company that builds agent-operated businesses ended up shipping verification infrastructure — and why the answer was sitting in a pattern the web has now had to learn twice — once from hacked certificate authorities, once from a poisoned software build…
About Steven
Steven Mih has founded and led four venture-backed companies, including Ahana — the cloud-native Presto managed service acquired by IBM in 2023 — and served as a board member of the Linux Foundation's Presto Foundation. Earlier: CEO at Aviatrix (cloud networking and security); sales and BD leadership at Couchbase and Mesosphere. Action State Group is the studio for the agent era: building the companies, and the trust infrastructure, that let teams operate far beyond their size. He serves on the board of the Francisco Park Conservancy in San Francisco.
1
Cloud Infrastructure
Built distributed systems giving developers massive leverage
2
Co-Founded Ahana
Cloud-native analytics company acquired by IBM (watsonx.data)
3
Co-Founded Across AI
Seed raised, team assembled, and customer engagements began. Continuing as Strategic Liaison (GTM).
4
Action State Group
AI studio: agent-operated businesses + open trust infrastructure for agent actions.
Build with us
We're looking for design partners running real businesses on agents, contributors to our open-source verification tools, and collaborators in the standards community.
Action State Group, Inc. · San Francisco · Open source: Apache-2.0 · verify.actionstate.ai