How Varick AI Uses FDE for Operational Transformation

These are my notes from the Forward Deployed Engineering workshop at @aiDotEngineer on June 30, 2026.

Session: AI Tools for Forward Deployed Engineering
Speaker: Vasuman Moza
Company: Varick Agents / Varick AI
X: @vasuman

Company Overview

Varick AI is an FDE firm focused on deep operational transformation.

The core thesis is that execution is no longer the bottleneck. Context and process understanding are. Generic AI calls break down quickly without rich business context.

Why FDEs Exist

Most AI pilots fail to reach production or measurable ROI. The root cause is that AI gets added to broken processes instead of redesigning them.

FDEs map workflows, redesign processes around AI, and decide what stays human.

Process Mapping

FDEs interview processes across AR, AP, card reconciliation, banking, billing, FP&A, and more. They document company-specific logic that is often undocumented.

Varick targets department-wide transformation, not point solutions.

• Point solutions yield 5% to 10% ROI.
• Varick targets 25% to 75% ROI.

Internal Tooling

JD Pruitt leads the FDE agent platform.

The platform was built with Codex. It includes a meeting assistant that pulls Granola notes and synthesizes documentation. It also includes an autonomous assistant for change requests, pending approval.

Varick has a model training layer and knowledge graph traversal tools for client-specific context.

My Take

Varick's thesis is that execution is no longer the bottleneck. The scarce thing is understanding how a business actually works.

That changes the FDE job. The engineer is not just plugging AI into an existing process. They are mapping the process, finding the hidden company-specific logic, deciding what should stay human, and then redesigning the operating model around AI.