Technical AI Strategist · Product Leader
I collaborate with teams to bring meaningful AI solutions to life.
20+ years of product leadership at Equifax, Global Payments, Unum, and Flock Safety — plus my own agentic AI products. I take contract and full-time engagements where an AI initiative needs an operator, not another strategy deck.
Google Cloud Certified — Generative AI Leader · SAFe 5 POPM · CSM · CSPO
Recommendations · via LinkedIn
What my peers say.
“I worked with Sam on the Product Innovation team at Equifax, and he was the guy constantly pushing us to stay ahead of the curve with modern tech and AI.”
“Sam has this great knack for taking messy, raw data and turning it into something visual that clients actually find useful. What he did with the modernization of Credit Trends and the Ignite® Marketplace was a huge win for the team. But he wasn't just about the tech; he handled a $25M risk portfolio with a lot of strategic grit and was always the one making sure Engineering, Legal, and Marketing were actually on the same page so our launches went smoothly.”
“Sam is one of those rare PMs who is super technical but also knows exactly how to navigate the "people" side of a big company.”
“I had the chance to work with Sam at Equifax, and he's one of those people you can rely on when things get complicated (which, as we both know, they often did).”
“He played a big role in pushing the Ignite product suite forward, not just from a strategy standpoint but actually getting things done. Sam has a good way of cutting through noise and turning ideas into something real and usable, especially when it comes to evolving products like Lost Sales Analysis and TradeSight. He helped move us away from manual processes into more scalable, real-time solutions, which was a big shift.”
“What I appreciated most was how he worked across teams. Whether it was pricing, marketing, or legal, Sam knew how to navigate the conversations and keep things moving without making it feel heavy. He also took the time to share what he was learning—especially around AI—which helped the broader team level up.”
“Beyond the work itself, he's just easy to work with. Steady, thoughtful, and always focused on moving things forward. I would call Sam a friend that I happened to work with. He is someone who was ahead of the curve with his passion for learning AI at a time when I was a novice and would lean on him to explain in layman's terms.”
“I'd absolutely work with Sam again and would recommend him to any team looking for someone who can handle both the strategy and execution side of the house.”
Selected work · 2001 → now
The work, told the way I actually think about it.
Not job descriptions — problems, approaches, outcomes. Metrics lead where they exist.
- ★ Flagship2022 — 2026EquifaxDirector of Product Management
Lost Sales Analysis — automated auto-lending intelligence
- Problem
- Auto lenders relied on slow, manual analytics to understand lost deals and refine decisioning criteria.
- Approach
- Designed and scaled an automated, real-time analytics product end to end — Voice of Customer discovery through launch — aligning Engineering, Pricing, Marketing, and Legal in a highly regulated credit environment.
- Outcome
- 2025 flagship launch, recognized by senior leadership. Lenders gained real-time insight to capture lost market share.
ai/analytics product · 0-to-1 · regulated industry
- 2022 — 2026EquifaxDirector of Product Management
TradeSight® for Automotive
- Problem
- Lenders and dealers lacked a comprehensive view of the dealer landscape and market-level opportunity.
- Approach
- Directed development of a market-intelligence product built on 80M+ auto loan tradelines, with trend analysis and opportunity mapping.
- Outcome
- Second flagship 2025 launch; full dealer-landscape visibility for automotive risk customers.
80M+ tradelines · data products · market intelligence
- 2022 — 2026EquifaxDirector of Product Management
Ignite® Marketplace Apps & Credit Trends
- Problem
- Raw credit data was underutilized; customers needed insight, not exports.
- Approach
- Managed a marketplace of analytic apps turning raw data into decision-ready visualizations; owned the Credit Trends behavioral risk product line.
- Outcome
- Measurable gains in customer satisfaction; SAFe 5 product frameworks adopted across global engineering teams.
data visualization · portfolio management · safe
- 2021UnumSenior Product Owner
Conversational AI for Leave & Absence
- Problem
- A $756M Leave & Absence business needed a scalable, user-friendly digital front door.
- Approach
- Spearheaded a flagship conversational AI web app; led multi-time-zone engineering and design teams; iterated NLP and dialogue management on NPS and customer feedback.
- Outcome
- Flagship AI product for a $756M business line — and a working bridge between data science and executive decisions.
$756M business line · conversational ai · nlp
- 2021 — 2022Global PaymentsSenior Technical PM
B2B2C commercial payments & APIs
- Problem
- Next-gen payment experiences — real-time installments, secure card management — needed 0-to-1 engineering leadership.
- Approach
- Orchestrated the full product lifecycle; translated complex requirements into prioritized sprint backlogs; cleared engineering bottlenecks.
- Outcome
- Accelerated release cycles for real-time installment and card-management systems.
fintech · apis · technical product management
- 2022Flock SafetySenior Product Manager
Operations through hypergrowth
- Problem
- Rapid company scaling threatened delivery quality and operational efficiency.
- Approach
- Partnered with revenue and field operations to optimize delivery cycles; orchestrated cross-department workflows.
- Outcome
- Product quality standards held through a rapid scaling phase.
operations · scale · cross-functional leadership
- 2014 — 2020CodeSlingEcommerce Product Manager
E-commerce platform modernization
- Problem
- Checkout throughput, deployment friction, and platform stability were limiting conversion.
- Approach
- Led cross-functional delivery of cloud features on Python, Azure, and GCP; guided stack decisions across JavaScript and Ruby.
- Outcome
- Improved checkout throughput, faster deployment cycles, better on-time delivery.
e-commerce · cloud · platform
- 2013 — 2014Tidbit.coCo-Founder & PM
Five-week mobile launch, Gigtank Accelerator
- Problem
- The hospitality sector needed mobile training tooling — fast.
- Approach
- Led a cross-functional team through an accelerator to build, launch, and fund a mobile app.
- Outcome
- Funded launch in under five weeks — a template for rapid 0-to-1 execution.
5 weeks, idea → funded launch · founder · rapid prototyping
- ★ A decade in GovTech2001 — 2013Georgia Superior Courts / GA PSCTechnical Services Mgr · CIO
Statewide judicial SaaS & agency modernization
- Problem
- State agencies ran on aging platforms with costly, fragmented data workflows.
- Approach
- Owned service delivery for statewide legal SaaS used by judicial staff; as CIO, rebuilt an agency technology platform, modernized data centers, and streamlined data collection.
- Outcome
- Reduced operating costs and a decade of trusted government technology leadership.
govtech · saas · enterprise modernization
Currentlybuilding AI products for home service operators and everyday businesses under the Amazing.ai brand.
Interlude
“Machines got me into this. Now I spend my days making the new ones behave.”
How I work
Good outcomes come from good operating systems.
These are the three frameworks I bring to every engagement.
01 / Process Improvement Framework
Find the friction. Fix it in order. Prove it stuck.
- Analyze. Find friction — delays, rework, confusing handoffs. Balance efficiency against effectiveness. Categorize work: predictable processes get lean (eliminate waste); complex, uncertain work gets agile (adapt to change). Map it with process maps and value stream maps.
- Implement. Plot problems on an Impact × Effort prioritization matrix. Start with quick wins, plan strategic initiatives, and fail fast with low-risk tests. Bring stakeholders along using the Power-Interest Grid.
- Measure & sustain. SMART KPIs tied to business goals; process metrics separated from business outcomes; regular check-ins and feedback loops so improvements don't decay.
ROI = (Time Before − Time Now) ÷ Time Before × 100
| Process Map | Value Stream Map | |
|---|---|---|
| Perspective | Internal: how do we do our work? | Customer: how does value flow to them? |
| Key data | Steps and roles | Steps, roles, and time — touch vs. wait |
| Main goal | Clarity and understanding | Speed and waste elimination |
02 / WAT — Workflows, Agents, Tools
Probabilistic AI reasons. Deterministic code executes.
That separation is what makes AI systems reliable.
- Workflows. Markdown SOPs that define the objective, inputs, tools, outputs, and edge cases — written the way you'd brief a teammate.
- Agents. The intelligent coordinator. Reads the workflow, sequences the right tools, handles failures gracefully, and asks clarifying questions. Connects intent to execution.
- Tools. Deterministic scripts for the actual work — API calls, data transforms, file operations. Consistent, testable, fast.
Why it matters: at 90% per-step accuracy, an AI handling five steps directly succeeds only ~59% of the time. Offloading execution to code keeps the AI where it excels — orchestration and decisions.
The self-improvement loop: identify what broke → fix the tool → verify → update the workflow → move on with a stronger system. Every failure makes the framework better.
Case: HITL lead enrichment
Token-optimized scraping strips HTML bloat before the AI sees it; PII never leaves the perimeter; a Slack “Claim Lead” button keeps humans in control of routing.
Case: support ticket triage
Deterministic keyword rules escalate outages instantly — no AI guesswork. A zero-risk policy means the AI can only write hidden drafts, never email a client.
03 / Spec-Driven Development (SDD)
Specs as executable contracts — not passive documentation.
Traditional development writes code first, docs later. SDD flips it: a rigid SPEC.md strictly guides and constrains what an AI agent is allowed to build. It stops agents from going rogue, inventing their own component styles, or shipping black-box code you can't maintain.
- Vision. Define the what and why — business requirements and constraints.
- Plan. The AI analyzes existing code read-only and produces a rigid SPEC.md blueprint before any code is written.
- Decompose. The spec is broken into tiny, isolated, testable task chunks.
- Execute & verify. Step-by-step execution while a validation framework checks code against the original spec — preventing architectural drift.
Supporting practices: multi-agent orchestration, golden eval tests, human-in-the-loop gates, and tiered AI routing — local-first models for high-frequency tasks, cloud escalation only for heavy reasoning.
this site was built this way.
Credentials & skills
The paper trail.
Certifications
- Generative AI LeaderGoogle Cloud Certified
- SAFe 5 POPMScaled Agile, Inc.
- Certified ScrumMasterScrum Alliance
- Certified Scrum Product OwnerScrum Alliance
- Claude Certified Architect FoundationsAnthropic Partner Academy · in progress
Competencies
AI product strategy & agents · agentic workflows & multi-agent systems · RAG & prompt engineering · AI governance & evaluation · cloud: GCP, AWS, Azure · Agile / SAFe program leadership
Education
- M.S., Information TechnologyAmerican InterContinental University · Atlanta, GA
- B.S.Oglethorpe University — Atlanta, GA
- Health Information Technology CertificateEmory University · Atlanta, GA
Let's build something reliable.
Recruiters and hiring managers: I move fast — send the req and I'll respond ASAP with fit, rate, and availability.
Current availability
- Contract (W2) or Corp-to-Corp via AmazingDotAi, LLC
- Full-time considered for the right remote and hybrid AI-focused role
- Remote (US) · hybrid Atlanta, GA / Greenville, SC / Chattanooga, TN / Nashville, TN
- Can interview immediately, start within 2 weeks
Tell me the problem — I'll bring the operating system to solve it.