Patents & Inventions
A patent isn’t just an idea — it’s a legal right granted because that idea is genuinely new. These are mine, co-invented with colleagues at IBM: real inventions in how AI can solve a problem in a new way.
Granted
Generative AI Intelligent Workflow for Augmented Support and Repair
U.S. Patent No. 12,619,933
Designed for the moment a technician is staring at something broken and doesn’t know where to start — a workflow that thinks alongside them, surfacing the right knowledge, the right procedure, the right next step, in real time.
Dynamic Maturity Model
U.S. Patent No. 12,705,562
Most maturity models stay stuck in their debut era — this one reinvents itself like a pop star. It reads documents and survey responses to build a live capability map across domains like data governance, culture and operations, then keeps rewriting the score as the organization changes.
Intelligent Secure Automation of Claim Preemptive Subrogation
U.S. Patent No. 11,449,951 B1
Why does an insurance claim happen after an accident when satellite data, dashcam footage, weather conditions, and driver history all existed at the moment of impact? This uses all of it — GPS, IoT sensors, witness accounts, jurisdiction-specific regulations — to de-bias fault assessment, settle the claim on a blockchain, and generate a plain-language explanation of how the decision was made. No adjusters. No dispute. Done.
Pending (Published)
Generative Artificial Intelligence-Enabled and Augmented User Research Using Synthetic Personas as Subjects
U.S. Pub. No. 2026/0127652 A1
Real user research takes real users — recruiting, scheduling, waiting weeks for insight. This builds AI personas instead, trained on real user profiles, product data, and company history, then scores how well a product or service actually fits the people it’s for. Wherever the fit is weakest, it flags that as the area to dig into — and runs the research on the synthetic users on the spot, no recruiting required.
Intelligent Workflow Event Prediction and Contingency Planning
U.S. Pub. No. 2025/0384291 A1
Most strategies break on contact with reality. An AI strategy engine detects exogenous disruptions — economic, geopolitical, market signals — generates divergent future scenarios using reinforcement learning, validates them with a human in the loop, selects the best path forward, and rewrites the operating model and benchmarks to match. Strategy and execution stay synchronized.
AI-Enhanced Intelligent Workflow for Improved Personal Performance
U.S. Pub. No. 2025/0335858 A1
Generic development plans aren’t actually about you. This one is — it benchmarks against peers and your own history, pinpoints exactly where the gaps are and why, then recommends specific projects, mentors, and calendar changes. Goals adjust as you progress. A career engine, not a career plan.
Dynamic Benchmarks of an Intelligent Workflow
U.S. Pub. No. 2025/0005396 A1
A strategy is only as good as what it measures — and most measurement systems are built once and never touched again. This is a closed-loop AI agent that continuously pulls from internal data and external signals, evaluates every benchmark against a quality metric, retires the ones that no longer fit, and introduces new ones. The strategy and its measurement system evolve together, automatically.
Intelligent Workflow for Maturity Assessment of Ecosystem-Enabled Innovation
U.S. Pub. No. 2024/0303573 A1
Most companies measure how good they are. This measures how good they are to work with — across partners, suppliers, and complementors. It never stops reading contracts, filings, and news, so as the ecosystem grows, so does the picture.
Intelligent Workflow for Healthcare
U.S. Pub. No. 2024/0144381 A1
Your doctor can only work with what they can see. This agent fills in the gaps — connecting data from your sensors and environment, flagging changes in your condition, recommending the right practitioners, and even suggesting the devices you need to get the information that matters. It can act on your behalf when you can’t, and a smart contract means you stay in control of all of it.
Distribution of Surplus Products Using Artificial Intelligence
U.S. Pub. No. 2024/0257953 A1
Volunteers in New York were matching surplus goods to people in need using spreadsheets and phone calls. There had to be a better way. This is a neural-network system that matches surplus goods to the people who need them and manages the supply chain to get them there — from cold-storing food until pickup to routing it through drivers along the way. It handles medicine, food, clothing, and consumables, anywhere in the world, at any scale.