Los Gatos & San Diego, CA

Jason Lee

Directing Agentic AI where it actually matters. At the constraints.

Scroll
The question is not which tasks can AI automate. The question is which constraints can AI eliminate.
— From "Beyond the Local Efficiency Illusion"

I'm a Forward Deployed Strategist at Google Cloud, where I help enterprises cross the gap between agentic AI pilots and production.

My work sits at the intersection of frontier models and enterprise reality — the messy, political, system-level place where most transformations stall. I've led engagements across regulated industries — clinical healthcare, financial services, manufacturing, defense — where the gap between an impressive demo and a deployable system is measured in months, not minutes.

I co-authored Google Cloud's Agentic Transformation Framework and write about why most AI investments fail to reach the P&L. I have a background in finance and behavioral economics, spent formative years working for international companies headquartered in Milan and New York, and currently split my time between enterprise strategy and red-teaming SOTA and Frontier Models with Google's AI Principles Pioneers.

I take on a small number of outside engagements each year. If any of the focus areas below match a real problem you're working on — not a vague curiosity — I'd like to hear about it.

Four areas where deep engagement changes the outcome.

01

Productionization of Agentic Solutions

Most agentic pilots stall on the way to production. The demo works, the steering committee applauds, and then the system meets real users, real data, and real compliance. I work that gap — the architecture, the evaluation discipline, and the unglamorous stakeholder choreography that turns a prototype into a system that can hold weight.

02

Multi-Agent Systems

Enterprise work isn't a pipeline — it's a weather system. Handoffs, approvals, parallel tracks, and exceptions that don't fit the spec. I design multi-agent architectures that respect that topology: orchestration that survives reality, guardrails that hold under load, and a clear theory of which agent is responsible for what.

03

Context Engineering

Agents are only as reliable as what they know. The difference between brittle and production-grade is almost never the model — it's the retrieval, the memory, the tool design, and the quiet discipline of finding where the system fails before your users do. I work on the context architecture that compounds into trust.

04

0–1 Frontier Innovation

The best opportunities don't live in the feature backlog — they live at the edge of what just became possible. I partner with founders and intrapreneurs on zero-to-one work where the hard question isn't "which model?" but "what shape is the thing we can now build that was science fiction six months ago?"

Papers, frameworks, and things I've built.

Partnerships I'm proud of.

A selection of publicly cited engagements I've contributed to at Google Cloud — largely in healthcare & life sciences, public health, and the systems that shape people's daily lives. The work I'm drawn to sits where agentic AI meets meaningful human outcomes.

2026

CVS Health × Google

Strategic alignment with CVS Health on the safe and ethical use of health AI — safeguarding patient data privacy, upholding responsible AI implementation, and building systems with equity as a cornerstone.
2025

HCA Healthcare × Google

Multi-workstream partnership across HCA's 190 hospitals — an AI-powered Nurse Handoff tool built with frontline nurses as co-designers (early pilots rated 86% factual, 90% helpful), alongside ambient documentation that converts clinician-patient conversations into draft EHR notes for physician review and approval.
2025

CME Group × Google Cloud Universal Ledger

Tokenization technology introduced to enhance capital market efficiency — a first-of-its-kind collaboration between the world's largest derivatives marketplace and Google Cloud's new Universal Ledger.
2025

Mayo Clinic × Google

Partnership pairing Mayo Clinic's clinical expertise with Google's cloud, data analytics, and AI — researching and building transformational tools that equip healthcare providers to improve patient outcomes at global scale.
2024

Taiwan NHIA × Google

Collaboration with Taiwan's National Health Insurance Administration on type 2 diabetes risk prediction and management — applying Google Cloud's AI and data analytics capabilities at national population scale.
2024

Dexcom × Google

Generative AI platform built with Dexcom — the global leader in continuous glucose biosensing — to help millions of users translate real-time glucose data into actionable decisions for their own health, powered by Google Cloud.
2023

McDonald's × Google Cloud

Strategic partnership to connect the latest cloud technology and apply generative AI solutions across McDonald's restaurants worldwide.

A working history.

2025 — Now
Forward Deployed Strategist · AI/ML
Google Cloud

Technical strategy for key enterprise accounts — transitioning customers from POC to production-grade agentic AI deployments across regulated industries. Codifying playbooks for 1-to-10 scaling.

2023 — Now
AI Principles Pioneer · Red Team & Advisory
Google Cloud

Red-teaming frameworks for Google's emerging generative AI models — in partnership with DeepMind and research teams across Core, Cloud, and Moonshot initiatives. Focus on fairness, harm, and human-rights risk.

2022 — 2025
Innovation Lead, AI Strategy
Google Cloud

Digital transformation strategy for Fortune 50 clients. Launched "Generative AI Innovation Factories" — end-to-end enterprise agentic solutions including Agentspace integrations, custom workflows, and RAG/GraphRAG knowledge assistants.

2019 — 2022
Transformation & Innovation Product Owner
Generali Group · Milan / New York

Global research and innovation lead across business lines. Built strategic frameworks and deployed technology opportunities across multi-cloud environments. Shaped the global product development roadmap.

2017 — 2019
Innovation & Strategy Consulting Partner
NDGE · New York

Lead strategist on competitive insights, market research, and product analytics. Built dynamic pricing models and reusable market research frameworks using conjoint and TURF analysis.

If you're working on something hard at the edge of agentic AI — I'd like to hear about it.

I take on a small number of 1:1 engagements each year — advisory sessions, constraint-diagnosis workshops, and 0-to-1 innovation sprints. If what you're working on matches one of the focus areas above, book a 30-minute intro below, or send a note directly.