AI in compliance

Compliance is often the slowest—and most frustrating—part of bringing a product to market. While product development has accelerated, meeting standards like ISO 27001 or SOC still takes weeks, delaying customer trust. Our solution brings this timeline down from days to hours by embedding AI into the heart of compliance readiness.

In this talk, I’ll walk through the two key user flows in our system:

For compliance-ready teams — Users upload their documents and select the compliance framework. Our AI agents collaboratively analyze these proofs against each requirement, reasoning over them with LLMs to identify gaps.

For founders new to compliance — A chat interface guides them step-by-step to meet the chosen compliance standard, making the process as simple as onboarding to a new tool.

We also address real-time compliance monitoring using MCP (Model Context Protocol) servers that connect to the user’s systems in read-only mode, collecting data, identifying risks, and suggesting fixes. We’re building an execution agent that will apply approved changes automatically.

Under the hood, we use:

Litellm as an LLM gateway to tap into the latest models (OpenAI GPT-5, Google Gemini 2.5 Pro, Claude, etc.)

Langfuse for agent monitoring

YAML-based prompt management for maintainability

LlamaIndex for document indexing to create a knowledge base from uploaded proofs

Our internally built compliance knowledge layer for ISO 27001 & SOC standards

ECS, Vercel, Next.js, and FastAPI for serving & UI

Attendees will see how LLMs, agent frameworks, and real-time integrations can remove compliance as a blocker, turning it into a continuous, automated process.