I lead AI transformation end to end: I turn work that people do by hand into software and AI systems, so a business grows without adding headcount.
At Croud I doubled revenue per employee and cut employee cost 15% by replacing manual delivery with software. Then I went further: I founded Titan, an AI operating system for multi-location service businesses. In the franchise locations where we proved it out, labor fell from 50% to 30% of revenue, and a seven-person engineering team became one person running AI agents. A national franchise group now licenses it across 120+ locations.
Every number that runs the day, in one place, and pushed to the people who need it before they go looking. No more digging through five systems to find out what happened.
Assistants trained on the company's own data and playbooks. Click a number to see why it moved, or ask how we do something, and get the company's answer, not a generic one.
The repetitive work comes off people's plates. Recruiting screens, invoicing, collections and follow-ups run as workflows, and people step in only for the calls that need judgment.
Errors, team feedback and feature requests flow back in and become fixes and new features, so the system gets better every week without a project to kick it off. How that works is below.
Where I start: find the one number that matters most, build the smallest system that moves it, prove it in production, then widen from there.
I built Titan with a software team made of AI agents. Every change goes through the same pipeline of reviews, tests and checks, and a person only steps in where a decision needs one. It's the AI transformation playbook I'd bring to a company that wants AI doing real work, not running pilots.
Each agent works in its own isolated copy of the code, and merges go through one queue, so work runs in parallel without collisions. I've run more than 80 at once. The limit isn't compute or headcount, it's how fast decisions get made, and the system is built to make most of them itself.
Plans and code are each challenged by AI reviewers for safety, architecture and test coverage, plus a reviewer built from 52 of my own review comments. Above them sits a CEO agent built on my judgment: it reviews the results, settles what's ready and pushes each workstream forward, so work doesn't wait on me.
Plans are published as Titan documents and reviewed inline before any code is written. Then come up to 21 automated checks and 400+ test files, a live preview on production-shaped data, a merge queue that refuses anything failing, and a deploy that tests itself and rolls back if it breaks.
Errors caught overnight, feedback from the team and new feature requests all flow back in. The system turns them into plans and code, runs them through the same reviews and pipeline, and ships the improvement. By morning, what broke is fixed and what people asked for is in progress. It emails me only when a decision truly needs a person.
Most of my career has been inside marketing: measurement, performance and growth for brands and the agencies that serve them. At Croud I worked with Chief Strategy Officer Avinash Kaushik to turn outcomes-first strategy into a practice teams around the world delivered the same way, and I built the framework the agency used to plan and measure client growth. It's why I approach AI transformation from the P&L, not the tech stack.
Read awareness, consideration, intent and conversion data together to find the part of the funnel with the most upside, and state the expected outcome before spending a dollar.
Plan for frequency, conversion and brand lift, and tie pricing and staffing to the client's results instead of hours billed.
Value-based bidding, propensity models and run / fix / kill rules that keep promoting what works and cutting what doesn't, creative and ad groups included.
Custom multi-touch and predictive attribution, brand-lift and incrementality testing, and customer lifetime value, so budget follows what actually drives growth.
In January 2023 I wrote that AI would commoditize agency services, and argued we should price on client outcomes and use AI to cut staffed hours. As VP of Data & Intelligence I then led that shift: AI-generated weekly reporting, a taxonomy tool that saved one client $900K, and revenue per employee doubled in a year.
My outcomes-based modeling was a major part of winning Croud's two largest clients. I sold analytics into all 15 of its top accounts and got a franchisor to roll my system out across its network. AI transformation fails on adoption, and adoption is where I'm strongest.
At Croud I took strategy engagements that depended on one person and made them a process teams around the world delivered the same way, and merged four teams into one operating model with clear roles, cadences and career paths. I don't stop at the framework, I build the product: I prototyped Croud's taxonomy builder, and with Titan I built the systems that turned recruiting, invoicing and collections from someone's job into software, using AI agents that work as extensions of me. An idea becomes a working tool people use in days, not a slide deck.
From TD Ameritrade's key moments to Pulse, Titan's daily dashboard and email, I design around the decision someone has to make: what to invoice, what to chase, where to send an open crew. When a number would mislead, the system holds it back instead of showing it.
I'm looking for a full-time role, remote from the Hudson Valley. The best next step is an intro to the team, or a short conversation about where AI is stuck in your organization.
Thank you. The most useful thing is a specific introduction. Any one of these takes a minute. If someone needs a file for their hiring system, here's my resume as a PDF.
Tick any where you could introduce me to the team or hiring manager.