Applied AI research and special projects. We build at the frontier — connecting intelligence to data, and data to decisions.
Applied AI research and special projects. We build at the frontier — connecting intelligence to data, and data to decisions.
The Lab Ethos
Every project in the lab connects to a real problem that organizations will face sooner than they expect. We don't theorize about what AI can do. We build it, measure it, and bring back what survives contact with reality.
Current lab work spans two active research areas — each with live production implementations and proprietary IP.
Lab Project 01 — Active
Connecting entity-resolved data with agentic intelligence
Most organizations have the data. What they lack is a system that knows who is in it — and can reason about them. Conversational Entity Intelligence combines the precision of entity resolution with the natural-language reach of agentic AI, producing systems that can answer questions about real people, relationships, and behaviors across fragmented data sources.
The research questions driving this work: What happens when you give an AI agent access to a properly resolved identity graph? What classes of question become answerable that weren't before? How do you design the interface layer between resolved entities and conversational intelligence?
Active Lab: P.O.E. — Platform for Obscure Experience
P.O.E. is deployed for Ad Obscura, LLC ("The Obscure") — a cutting-edge distillery and entertainment company based in Los Angeles, CA. P.O.E. helps The Obscure understand who their customers are, how they are related to one another, what they are doing, and what actions the business can take next to enrich the customer experience — all through live conversational intelligence.
Data in scope
Capabilities
The Obscure stands apart from all other craft distillers and entertainment companies as an AI Native Company — building intelligence into the guest experience from the ground up.
Visit The Obscure in LA →Lab Project 02 — Commercially Validated
Rapid Opportunity Observation & Targeting
Feature-driven backlogs optimize for output. ROOT is built to optimize for outcomes — replacing static roadmaps with a continuously updated, evidence-directed investment model that governs complex programs from inception through long-term operations.
At its core, ROOT treats product discovery as a data problem. Every signal — qualitative or quantitative — is ingested, labeled, and mapped to a live knowledge graph of opportunities, constraints, and value levers. Investment decisions become auditable, adaptive, and grounded in real evidence rather than assumption.
ROOT is domain-agnostic. It applies wherever business outcomes can be defined and data exists to surface constraints against those outcomes.
Governing Principle
Evidence-Based Innovation — every investment decision traceable to observed data, every roadmap adjustment driven by what was learned, not what was assumed.
The Six-Phase Framework
Core System
OST Knowledge Graph — dynamic, traversable, 4-tier opportunity structure with live telemetry feedback
Interface
Conversational agentic layer — stakeholders query the evidence base and opportunity graph in natural language