Building the Trust Layer for AI Agents - Technical Co-founder

AI agents are moving from answering questions to taking action inside real business workflows.

Soon, AI systems will follow up on invoices, prepare refunds, handle supplier requests, review contracts, touch customer data, route approvals, and trigger financial workflows. But serious companies will not let AI agents act without control.

Before AI can operate inside finance, procurement, contracts, or operations, companies need to know: what was the agent allowed to do, what evidence did it use, who approved it, and what happened afterwards?

That is the problem Relay is solving.

Relay is building a trust and control layer for AI agents: a structured approval object for AI-generated work, containing the proposed action, evidence, permissions, approvals, and audit history before anything is executed.

The first wedge is invoice exceptions in finance teams: overdue invoices, payment disputes, "already paid" claims, invoice-copy requests, and messy customer follow-ups.

Instead of letting AI agents act freely, Relay helps finance teams use AI safely - faster workflows, less manual work, and clear control over every AI-prepared action.

Long term, the vision is bigger: to become the default control layer for delegated AI work.

Current stage and proof

This is early, but not just an idea. We have built an initial MVP, mapped the product/system architecture, started customer discovery with finance/accounting users, receiving support from KTH Innovation, and validated early pain points with people in finance, accounting, and business operations around approvals, auditability, and AI-driven workflows.

The next step is to turn these insights into a focused alpha that finance teams can test. The technical challenge is to connect AI actions with business context, evidence, permissions, risk rules, human approvals, and a reliable audit history while integrating with existing company systems.

Relay is led by Klara Ridder, an Information Technology student at KTH with a background in strategic communication, product, and startup strategy. We are now looking for a Stockholm-based technical co-founder who wants to build the product and company from the ground up.

We are looking for someone who:

• has strong full-stack or backend engineering ability
• can build APIs, databases, authentication, and modern web apps
• is excited about AI agents, automation, infrastructure, and secure systems
• cares about product, users, and real business problems
• moves fast, builds scrappy prototypes, and improves based on feedback
• wants ownership, responsibility, and meaningful impact

Nice to have
Experience with Next.js, TypeScript, Python, Supabase/Postgres, AI/LLM integrations, MCP, workflow automation, audit logging, permissions, identity, fintech, accounting, procurement, or enterprise software.

What is offered
This is an early-stage co-founder opportunity with both risk and upside. You will help shape the technical architecture, product direction, and company.

The project is currently pre-funding, but the intention is to build Relay together with shared equity.

Initial commitment can start alongside studies or work, as long as ambition, pace, and ownership are clear.

If this sounds interesting, reach out with what you have built and why this space interests you.