Blog

Building AI agents
with UI.

How we deploy custom AI apps with purpose-built interfaces for both AI agents and human underwriters.

Vision Mar 31, 2026

The Runtime-First Approach to AI: Why Architecture Order Matters

Why we built the deployment runtime before we built the AI. And why that order changes everything about how humans and agents collaborate.

Read more →
Agentic UI Mar 31, 2026

Designing UI for AI Agents: Why Black-Box Models Fail in Commercial Insurance

Insurance is built on trust and auditing. An AI that just outputs a quote is a black box no underwriter will trust. To make AI work in insurance, the AI needs its own User Interface.

Read more →
Architecture Mar 30, 2026

Multiplayer Underwriting: Building a Real-Time Workspace for Humans and AI

What if an AI agent wasn't a background script, but a coworker editing the same document at the same time? That's how Oyyo's multiplayer workspace works.

Read more →
Engineering Mar 28, 2026

Why OCR Failed Insurance (And How LLMs Actually Solve ACORD Parsing)

Traditional OCR promised to automate insurance document processing for decades. It never delivered. Here's why LLM-based processing is fundamentally different.

Read more →
Product Mar 28, 2026

The End of Low-Code: Dynamically Generating Custom AI Apps for Specialty Insurance

Specialty insurance is too niche for off-the-shelf software. Building custom portals takes 12 months. What if AI builds the software for you in 30 seconds?

Read more →
Design Mar 26, 2026

Three Interfaces, One Workflow: Designing UX for Brokers, Underwriters, and AI Agents

The biggest mistake in enterprise AI is forcing everyone to use the same interface. Brokers, underwriters, and AI agents each need their own UX.

Read more →
Vision Mar 25, 2026

Mostly Autonomous, Entirely Human: The Future of the Underwriter

We're not replacing underwriters. We're giving them superpowers. Here's our vision for how insurance will work in 2035.

Read more →
Architecture Mar 24, 2026

Orchestrating the Ensemble: How Multiple AI Agents Auto-Quote a Policy

A single 'God Model' prompt fails at complex tasks. To auto-quote a commercial policy, you need an ensemble of specialized micro-agents working in relay.

Read more →