In progress · Jul 2026 – present · Founder / Lead Engineer
An AI-native operating system for founder-led teams
Overview
One governed platform where specialised AI agents work across a small business: websites and CMS, CRM and communications, newsletters, analytics, integrations and an installable skills marketplace.
The problem
Small teams stitch together disconnected tools for websites, customer communication, content, analytics and operations, while generic AI assistants lack safe access to business context and actions.
The solution
Pairly gives each business one governed platform where specialised agents can work across those functions. It combines tenant-scoped AI agents, websites and CMS, CRM and communications, newsletters, analytics, integrations and a skills marketplace.
What I built
- Tenant-scoped agent runtime with durable memory, model routing, typed tools and permission ceilings
- Streaming agent UX with human approval gates, audit events and token budgets
- Installable skills marketplace with automated validation and human publication review
- End-to-end features spanning API, CMS, renderer, email and PDF output
Solved ahead of time
Designed in, not bolted on.
Agents that can't overreach
Permission ceilings and typed tools bound what each tenant's agents can do.
Humans approve consequential actions
Approval gates and audit events wrap anything an agent does that matters.
Cost control
Token budgets are part of the agent runtime from the start.
A safe marketplace
Skills pass automated validation and a human publication review before anyone can install them.
Infrastructure
The choices. And why.
- PostgreSQL
- Durable state, with row-level security for tenant isolation.
What success looks like
A broad working platform with multi-tenant permissions, real-time agent execution, publishing and revision workflows, provider integrations and installable skills, with approval gates, audit events and human review for consequential AI actions.
Ask the twin