Inspiration
Triage started with a simple observation: high-level business intelligence is a luxury usually gated behind expensive consulting firms. For a small business or a startup, "due diligence" often means an owner sitting with ten browser tabs open, trying to make sense of conflicting data and hoping they didn't miss a red flag. We built Triage to democratize this process, turning hours of scattered research into a synthesized, affordable risk report. While it’s robust enough for a large enterprise, our heart was in making professional-grade consultation accessible to the businesses that usually can't afford it.
Looking at the Harper Personal AI Agent track, we saw a perfect opportunity to evolve this. We realized that true accessibility isn't just about providing a report; it's about automating the actual business workflows and decision-making that follow. By aligning with Harper’s mission of building intelligent agents for enterprise workflows, we’ve developed Triage to act as a tireless, "AI-native" consultant that works for everyone from the local shop to the global corporation.
What it does
Triage orchestrates a multi-agent pipeline to synthesize findings into a unified, confidence-weighted risk assessment across five critical dimensions: Due Diligence, Stress Testing, Competitive Intelligence, Compliance, and a final Master Report. It is an autonomous AI-powered due diligence agent designed to transform weeks of manual research into a streamlined, high-stakes risk analysis for modern businesses. By simultaneously querying a diverse array of intelligence sources, including public records, web data via Tavily, and internal documentation (in the future), the system aggregates fragmented information into a single, cohesive intelligence stream. Rather than presenting raw data, this approach helps detect deep-seated red flags, like financial instability or regulatory non-compliance, while maintaining transparency through explainable reasoning.
How we built it
“I used the stones to destroy the stone” -Thanos
Challenges we ran into
Lack of sleep
Accomplishments that we're proud of
Working MVP
What we learned
Building for the Harper track at YHack taught us that in the world of Enterprise AI, context is everything. An agent is only as good as the specific data it can fetch and the "reasoning" it applies to that data. We learned that enterprise-ready agents need to be more than just chatbots; they need to be reliable, transparent, and capable of handling structured data pipelines that a human can actually trust for high-stakes decisions.
What's next for Triage
Our goal is to evolve Triage from a "point-in-time" research tool into a fully integrated, proactive business partner. The next step is to fully integrate directly with "business-as-usual" tools like email, calendars, and meeting notes. By staying embedded in these daily workflow channels, Triage will act as a real-time monitor, automatically flagging risk points, regulatory shifts, or competitive threats as they emerge in your conversations and scheduling. We’re moving from a dashboard you have to visit to an intelligent layer that watches your back 24/7.
