Leads AI transformation across six enterprise functions · Seattle
Turning AI strategy into operating systems.
Head of Agentic AI & Automation, across Product, Engineering, QA, Support, Customer Success, and Professional Services. Founding member of an enterprise AI Center of Excellence advising 300+ business units.
What I lead
Depth across functions, not a pilot in a corner.
Agentic AI and automation across six functions
Product, Engineering, QA, Support, Customer Success, Professional Services. Discovery, prioritization, prototyping, governance review, production readiness, adoption.
Leads the function across those six areas, from discovery through production readiness, and coaches eight professional services teams on solution design, workflow fit, feasibility, tool selection, and output quality.
Founding member of an enterprise AI Center of Excellence
300+ business units across a global software portfolio. Built the intake, prioritization and value-realization frameworks the group runs on.
Enablement adopted by 40+ business units through executive workshops, role-based learning, AI literacy programs, and champion networks.
Fifteen years leading service organizations
Cybersecurity SaaS, regulated healthcare SaaS, enterprise fleet software. Ten of those applying AI, starting with Salesforce Einstein and Service Cloud.
Previously hired, built, and led a 16-person organization.
The discipline
Four gates separate a scalable bet from an experiment.
Most organizations do not have an AI idea shortage. They have a portfolio discipline and accountability problem.
Outcome
What measurable result will change?
Owner
Who is accountable for realizing it?
Workflow
Does the complete process improve?
Governance
What autonomy and risk are acceptable?
If an AI initiative has a model but no measurable outcome and no business owner, it is still an experiment.
Adoption
The gates decide what gets built. SPARK decides whether anyone uses it.
Most AI initiatives do not fail a technical test. They pass one and then never change how the work gets done. SPARK is the adoption framework I use on the other side of the gates.
Start small
One workflow where the pain is already measurable.
Pilot
Low risk, real work, where mistakes are easy to catch.
Analyze
Against the baseline, not against expectations.
Refine
Fix the workflow, not just the prompt.
Keep growing
Champions carry it further than a central team can push it.
Decision training
The four gates, playable.
Two simulators built for corporate training on AI adoption and transformation — putting leaders and stakeholders through the decision before they have to make it for real. Free, no sign-up.
Greenlit
Outcome and Owner
Eighteen AI pitches land on your desk. Twelve months, five points of engineering capacity. Most of them should not be built. Which ones do you sign?
Six verdicts. Teaches that transformation means saying no.
Play GreenlitRollout
Workflow and Governance
Lead an AI transformation at NimbusCare. Sixty weeks, a trust meter, and a clock. Can you get a company to adopt AI before time runs out?
Teaches that adoption is a people problem.
Play Rollout“The six verdicts create a strong framework for separating genuine AI opportunities from automation, process improvements, and initiatives that should be deferred or stopped.”Michael MoserCEO and Managing Director — on Greenlit
“It’s a lesson plan and a hiring tool in one, a real way to see if a leader can navigate those tradeoffs under pressure instead of just talking about them in an interview.”Lynn WalkerGrowth Marketing Executive, SLW Growth Marketing — on Greenlit
“An easy way for everyone to grasp the concepts pretty quick, and identify gaps for improvement.”Pad KankipatiAI/ML Architecture and Technology Executive — on Rollout
Method
How the work actually gets done.
Not every problem needs AI
Automation, a process change, or an AI assist is often enough. Reaching for a full AI solution when a simpler one clears the bar is how budgets get spent without the workflow improving.
Match the solution to the problem, then justify the difference.
Specify before you build
When code gets cheap, clarity gets expensive. AI does not remove bad assumptions. It industrializes them.
Tests validate what we expected. Personas expose what we forgot. Domain experts catch what neither the specification nor the model understands.
Cognitive diversity over consensus
Multiple frontier models on one problem, each prompted to refute rather than review.
Two models agreeing for the same wrong reason is correlated error, not corroboration.
Human judgment stays inside the system
The workflow gathers context, supports a decision, and routes uncertainty to a person.
Autonomy is a setting, not a goal.
Published
What I put in the open.
The same enablement work, without the enterprise wrapper. Free, MIT licensed, and tested in real use before it goes up.
The Cowork Blueprint
A practical blueprint for configuring Microsoft 365 Copilot Cowork for real work. Tested prompts, custom skills, playbooks, and templates — including the Compass: cloning your own thinking into the assistant so its output carries your judgment instead of generic AI voice.
“Not a feature overview or a product review. A working notebook from someone configuring this day to day, sharing the patterns that hold up and the ones that do not.”
Read the blueprintThe four plugins are the multi-model check-and-balance stack behind cognitive diversity over consensus — several frontier models on one problem, each prompted to refute rather than review.
Speaking
Two keynotes, and a standing invitation.
- UC Berkeley Executive EducationGuest lecture and fireside chat
- SEE 2026Grapevine, Texas
- EDGE 2025Oceanside, California
- Keynote, CopenhagenAnnual portfolio conference, 2026
- Keynote, CancunSame conference, 2024
Topics: moving AI from experiment to operating system · operationalizing agentic AI platforms · frontline adoption and building trust · the executive four-gate playbook · knowing when not to use AI.
Get in touch
Conversations I am always open to.
- Conference talks, panels, and fireside chats
- Executive education and workshop sessions
- Advisory conversations on AI adoption, governance, and operating models