Matt Phillips seated on stage with a microphone. The screen behind him reads Matthew Phillips, Director of AI Strategy, Asurion.

Hello, I'm

Matt Phillips

Director of Intelligent Engineering & AI Strategy

Artisan Studios

AI Strategy · Agentic Workflows · Human-Centered Transformation

About

From AI Pilots to Production Value

AI operator with a decade of data and AI product leadership, shipping AI-enabled products and workflows at scale in regulated, high-complexity environments. I take initiatives from problem identification through deployment to measured ROI. Not building decks, building outcomes.

Today I lead the GenAI practice at Artisan Studios, a technology consultancy, where I build repeatable AI delivery frameworks and prototype hands-on with Claude Code, AWS Bedrock, and agentic workflows. Before that, I spent 14 years at Asurion growing from product marketing into leading AI/ML product, generative AI strategy, and enterprise AI transformation for a business serving 300M+ consumers.

"People thrive on purpose, not pressure."

10

Years in data & AI product leadership

$12M

Documented value from a secure on-prem AI platform

133%

Of AI adoption target via an embedded Champions network

600+

Employees in the internal GenAI community I built

My Journey

From Marketing to Machine Learning

  1. Mar 2026 to Present

    Director of Intelligent Engineering & AI Strategy

    Artisan Studios

    Building and scaling the GenAI practice at a full-service technology consultancy, with client work across Government, Food Service, and Supply Chain. I architect repeatable AI delivery frameworks that compress time-to-pilot for non-technical operators, prototype directly with Claude Code, AWS Bedrock, AgentCore, and agentic workflows, and lead vendor evaluation for client AI initiatives, from scorecards and live-data pressure tests to AI-specific contract terms.

  2. Nov 2024 to Mar 2026

    Director of AI Transformation & Strategy

    Asurion

    Led enterprise AI adoption and change management across a $10B+ subscription business serving 300M+ consumers through 50+ carrier partnerships. Scaled an AI Champions network to 133% of its adoption target, built a Responsible AI governance framework aligning 20+ stakeholders across security, legal, and operations, and shipped AI voice simulation for sales training (23% sales lift for the pilot team) and IVR AI interactions (43% increase in call containment).

  3. Mar 2023 to Nov 2024

    Director of Generative AI Product Strategy

    Asurion

    Launched Asurion's first Expert Copilots, LLM-powered tools for technical support agents, and designed the LLM evaluation framework used to benchmark frontier models on latency, cost, accuracy, and task fit. Presented AI strategy and future-of-support prototypes to the Board of Directors, unlocking investment in GenAI infrastructure.

  4. Sep 2019 to Mar 2023

    Director of AI/ML Product Management

    Asurion

    Led AI product managers, ML engineers, and data scientists to scale real-time speech-to-text from prototype to production with a 55% accuracy improvement, securing $2M+ in investment that became the foundation of Asurion's AI services. Championed a secure on-prem AI cluster that generated $12M in documented value over two years, and established Asurion's first Technical PM and ML Engineering career tracks.

  5. Oct 2017 to Sep 2019

    Senior Product Manager, Analytics

    Asurion

    Rebuilt knowledge base search across 30+ carrier partners, cutting call handling time by 27 seconds and saving $1M+ annually. Launched an internal data education platform with 150+ courses and 200 recurring users, and implemented same-day device delivery reaching 85% of the U.S. within a 50-minute drive.

  6. Sep 2012 to Oct 2017

    Senior Product & Marketing Manager, Client Solutions

    Asurion

    Evangelized and commercialized a new messaging platform that delivered $58M in EBITDA, and drove 1.5M subscriber net adds, 20% app adoption, and a 15% churn reduction through partner marketing across retail, customer care, and digital.

  7. Earlier career

    Earlier career

    Cincinnati Bell

    Product and partner marketing in wireless, managing the prepaid business and helping launch early smartphones with OEM partners, which grew into a career-long focus on turning new technology into customer value.

What I Bring

Strategy That Ships

AI Strategy & Roadmapping

Turning ambiguous AI mandates into prioritized roadmaps that product, technology, security, legal, and operations leaders sign off on.

  • Roadmapping
  • Prioritization
  • Executive Alignment

Proof Board-level AI strategy presentations at Asurion

Agentic Workflows & Hands-On Prototyping

Bridging ideation and deployment by building working prototypes myself instead of waiting on engineering capacity.

  • Claude Code
  • AWS Bedrock
  • Agentic Systems

Proof AI-native delivery pattern from idea to deployed app

Read: From Bench to Building: My AI Delivery Pattern

LLM Evaluation & Model Selection

Benchmarking frontier models on latency, cost, accuracy, and task fit so investment decisions rest on evidence.

  • Evaluation
  • Benchmarking
  • Model Strategy

Proof Built the evaluation framework behind Asurion's GenAI investment decisions

AI Vendor & Partner Evaluation

Separating demos from durable value with scorecards, live-data pressure tests, and AI-specific contract terms.

  • Scorecards
  • Build vs. Buy
  • Contracts

Proof Vendor evaluations at Asurion and for Artisan clients

Focus

AI on Purpose

AI isn't failing because of the technology. It's failing because of how we introduce it to people. Three principles shape how I lead, build, and write.

HUMAN

Purpose over pressure

Mandates and fear narratives produce shallow adoption. People engage deeply when they understand why AI matters to their work and are trusted to grow into it. Fluency is built through repeated use on real problems, not completed training modules.


How I relate to this

At Asurion I scaled an embedded AI Champions network to 133% of target and grew an internal GenAI community from a handful of people to 600+ employees, by bringing people along rather than pushing them.

Read: When it comes to AI transformation, people thrive on purpose, not pressure.

TRUSTED

Constraints are a feature

Agents are built for completion, but enterprises need judgment. The tools that win will be the ones risk and compliance teams can approve: agents that understand the constraint layer around a user's intent and know when to stop and ask.


How I relate to this

I built Asurion's Responsible AI governance framework, aligning 20+ stakeholders on principles, risk standards, and deployment guardrails before AI reached production.

Read: The Next Real Differentiator in Agentic AI Isn't Capability. It's Refusal.

PRACTICAL

Build to learn

The barrier to building isn't knowledge anymore. It's willingness to stay in the problem. Leaders who prototype make better calls about timelines, risk, and what's actually possible.


How I relate to this

I use a repeatable chain: chat to frame the problem, a work agent to produce the plan, and Claude Code to ship it. It's how I go from idea to deployed software, and how I set realistic expectations with clients.

Read: From Bench to Building

Speaking

On stage at GenAI Week

The New Customer Service Playbook: AI-led with Human Oversight & Governance Systems

Event
GenAI Week Silicon Valley 2025, Santa Clara, July 2025
Format
Fireside chat with Michael Dimler, SVP Support Operations, Asurion
Role
Director of AI Strategy, Asurion
Topics
GenAI value realization, change management, and customer experience as a competitive edge

Read the LinkedIn post

Photos by Kevin Mei

Hands On

What I've Built

Claude Code engineering plugin

A Claude Code plugin that gives a delivery team one shared engineering method: 5 skills, 3 review subagents, 5 guardrails, and data schema reference SOPs. Engineers install it with two commands, and updates ship as reviewed, versioned PRs from a single owner.

  • Structured PR review with a refute-before-report pass. A separate subagent tries to disprove each blocking finding before it is reported, which cuts false alarms and ends in one clear merge verdict.
  • Spec-driven development, automated end to end. The plugin drafts architecture decision records (ADRs) and implementation specs, checks them against the live codebase before writing, and runs a four-pass review: a scripted template check, a fresh-context grounding check, a consistency check, and a hostile reviewer pass.
  • Claude Code

Intelligence Engineering Library

A ten-page library covering the full AI delivery path, from spec-driven development through pre-deployment evals to production observability, giving engineering teams one shared standard for building AI systems.

  • Onboarding guides for Cursor, Claude Code, and Claude Code plugins, with a cross-tool mapping so engineers can switch tools without relearning workflows.
  • Promptfoo as the pre-deployment quality gate and Langfuse as the production monitor, including a layered anti-hallucination approach for RAG.
  • AgentCore 101 and AgentCore Skill 101, covering CLI workflow, Cedar policy, evaluation, and a production setup checklist.
  • Claude Code
  • Cursor
  • AWS Bedrock AgentCore
  • Strands
  • Promptfoo
  • Langfuse
  • Confluence

World Cup Pick'em League

A family-and-friends prediction game with group-stage picks, a two-phase knockout bracket, a Golden Boot podium, match simulation, an admin panel, and automatic leaderboard scoring. Went from a walk to a deployed site in about a day.

  • Claude Code
  • Vercel
  • Supabase
  • TypeScript

This site

Built from a written spec with an AI coding agent, sourcing articles from Substack automatically.

  • Astro
  • Tailwind
  • Vercel

Testimonials

What Colleagues Say

Insights

Notes From Inside the Work

Start here

What Is This Crazy Place

Why I started writing: an internal GenAI community that grew to 600+ people, and a belief that AI succeeds when we bring people along with us.

  • Start Here
  • Community

From Bench to Building (Part 1)

A career spent adjacent to building, and the moment that changed. The barrier to shipping software is no longer knowledge, it's the willingness to stay in the problem.

  • Career
  • Builders
  • Mindset

InsightsSubscribe on Substack

Credentials

Always Learning

Education

  • B.S. in Marketing, Northern Kentucky University

Featured certifications

Anthropic

AI Fluency: Framework & Foundations

Jul 2026

ID x44ray26it2v

Anthropic

Claude Code in Action

Apr 2026

ID gzsdrd9spq7m

DeepLearning.AI

MCP: Build Rich-Context AI Apps with Anthropic

Aug 2025

Perplexity

Perplexity AI Business Fellowship

2025

Maven (Allie K. Miller)

AI for Business Leaders

Aug 2024

ID WgpFMyoW

Udacity

AI Product Management (Nanodegree)

Additional certifications
  • AI First Academy (Allie K. Miller / Open Machine)
  • AI Proficiency and AI Data Analysis, Section
  • Collaborative Writing and Coding with OpenAI Canvas, DeepLearning.AI, Dec 2024
  • AI Python for Beginners, DeepLearning.AI, Jan 2024
  • Bystander Training: From Bystander to Upstander, LinkedIn, Jun 2022
  • Nir Eyal on Creating Habit-Forming Products, LinkedIn, Nov 2020
  • Data Visualization: Storytelling, LinkedIn, Oct 2020

Awards

  • 2x Leadership Award, Data & Analytics Org (Q4 2021, Q3 2023)
  • 2nd Place, Asurion Hackathon 2025
  • Top Partner Marketing Team (2015)

Speaking & Media

Sharing What Works

Signature topics

  • Purpose over pressure: leading AI adoption without fear
  • Constraint-aware agents: governance as a product feature
  • From pilot to production: making AI pay off in regulated operations
  • The builder-leader: why AI strategists should prototype

Contact

Let's Talk

I'm always up for a conversation about AI strategy, agentic systems, or bringing teams along through AI transformation.

Email me