Company · About

The AI Security Runtime company

First Recon AI builds AI Security Runtime. The platform understands, secures, and proves every AI interaction, from device to model, so enterprises can adopt AI quickly and govern it with confidence.

Why we built First Recon

First Recon AI emerged from internal R&D at Persefoni, where we applied large language models across our own business. We were our own first customer: a regulated enterprise platform working out how to put AI to work without losing control of its data.

As we evaluated the AI security tools available, the gap was plain. Most were not designed for how enterprises adopt, govern, and secure AI. They rest on perimeter assumptions and bolt-on policies that break down once AI is in daily use, and they cannot read a prompt, judge the intent behind an interaction, or act before data reaches a model. Having led large IT, data, and security organizations ourselves, we knew secure adoption takes more than add-on tooling.

The technical foundation came from the same work: lightweight, quantized language models efficient enough to run security enforcement locally, on the device where AI use happens. That research set the runtime's enforcement position: policy applied inline, before the prompt reaches the model's API. It pairs on-device enforcement with centralized inspection, governance, and audit in one platform. We built the layered controls we could not buy.

PERSEFONI R&D APPLYING LLMS TO OUR OWN BUSINESS ON-DEVICE ENFORCEMENT QUANTIZED MODELS AT THE ENDPOINT FIRST RECON AI THE AI SECURITY RUNTIME COMPANY BLACK HAT USA 2026 LAS VEGAS · AUGUST 1–6
Foundation

The proving ground

500+Persefoni customers
8,000+Organizations · Persefoni
$179MRaised · Persefoni
PhoenixHeadquarters

Figures are Persefoni's: First Recon AI was built inside Persefoni, whose platform serves The New York Times, Snowflake, Xerox, Apollo Global Management, and SMBC.

What carries over is the operating discipline, not the history. Persefoni runs a regulated, enterprise-grade platform that has passed the security and compliance reviews of some of the world's most demanding enterprises, and First Recon AI is built with the deployment models, support rigor, and review scar tissue that operation earned.

Our own security practices are documented, in the open, on the security page.

Founders

The founding team

First Recon AI was founded by the team behind Persefoni, who have spent years operating an enterprise platform under the scrutiny of large, regulated customers.

Kentaro Kawamori

Kentaro Kawamori

Co-Founder & CEO

Also co-founder and CEO of Persefoni, where First Recon AI was built. Former Chief Digital Officer at Chesapeake Energy, a Fortune 500 company, where he led IT, digital, and data. Co-founded Umbrage, acquired by Bain & Company; Forbes 30 Under 30; previously Cloud Practice Lead at Accenture.

Kim Stroh

Kim Stroh

Co-Founder

Co-founder and CTO of Persefoni, and a computer scientist and mathematician with deep product and architecture expertise. A decade leading proprietary application development at a Fortune 500 company; an enterprise SaaS pioneer.

Jason Offerman

Jason Offerman

Co-Founder

Co-founder and President of Persefoni, with nearly 30 years across operations, strategy, and engineering. Led one of the most significant operational technology rollouts in the energy sector; advisor at Rice Investment Group.

Approach

How we build

BUILT FROM THE INSIDE OUTWe have run enterprise IT and security organizations; we build for constraints we know firsthand
SECURITY-FIRST, PRODUCTIVITY-ENABLEDEnterprises should put AI to work without giving up control, governance, or compliance
LAYERED, NOT BOLTED ONReal AI security is defense in depth: on the device, at the model boundary, and across AI traffic
ENTERPRISE-PROVEN FOUNDATIONSBuilt on the operating discipline of a regulated enterprise platform, backed by institutional investors

Today, AI Security Runtime is delivered as both a secure AI workspace and an endpoint agent, powered by the Semantic Security Engine™ and the Security Context Graph™. It secures and governs how enterprises use AI, and gives security teams observability, control, governance, and compliance across their enterprise AI.

First Recon AI is headquartered in Phoenix, Arizona, launches at Black Hat USA 2026 in Las Vegas, and partners with Conscia Group, a pan-European cybersecurity and managed IT services provider, to serve European enterprises.

Secure every
AI interaction.

30-day free trial