Prefactor Pty Ltd
Prefactor evaluates, observes, and improves AI agents in production in real time across every framework and workflow.
At a Glance
- Enterprise AI Teams
- Platform Engineering
- MLOps
- Security & Governance
AI Tools by Prefactor Pty Ltd
(1)
Prefactor
AI Agent Runtime Enforcement SDK
Discussions
No discussions yet
Be the first to start a discussion about Prefactor Pty Ltd
Latest News
Building observable agent workflows that humans can review at speed
From Lab to Liability: Why Agent Benchmarks Fail in Production
Prefactor Launch on Product Hunt
Announcing Prefactor’s Integration with OpenClaw
Products & Services
Real-time evaluation, observation, and improvement platform for AI agents in production. Includes agent registry, live tracing, and cost tracking.
Lightweight TypeScript and Python SDKs for clean integration at the agent runtime layer.
Command-line interface for managing evaluations and agent versions.
Market Position
Prefactor differentiates by operating at the agent runtime layer with real-time scoring and enforcement, unlike retrospective observability tools.
Leadership
Founders
Matt Doughty
Co-founder & CEO at Prefactor. Focused on bridging the gap between security and engineering for AI agents. Previously worked on observability and evaluation solutions.
Simon Russell
Co-founder & CTO at Prefactor. 25+ years of engineering expertise. Previously Founder & CTO at Overspecific, CTO at Tokio, CTO at iflix, and leadership roles at Greensync and Antler.
Executive Team
Matt Doughty
Co-founder & CEO
Experienced in observability and bridging the gap between security and engineering teams.
Simon Russell
Co-founder & CTO
Veteran engineering leader with experience scaling platforms like iflix and Overspecific.
Founding Story
Prefactor was founded to bridge the 'evaluation gap' between AI agent demos and production deployment by providing real-time visibility and runtime controls.
Business Model
Revenue Model
Usage-based subscription model based on the number of 'spans' (steps an agent takes).
Pricing Tiers
25k spans a month included; scored and risk-checked live.
100k-400k spans included; 100% of activity in real time; unlimited seats.
4M+ spans a month; SSO, audit retention, SLA, and named engineer.
Target Markets
- Enterprise AI Teams
- Platform Engineering
- MLOps
- Security & Governance
- Enterprise AI agents in production
- Compliance auditing for autonomous workflows
- Real-time risk monitoring
- Cost and performance attribution
- Versent
- Bupa
- CAR Group
- University of Sydney