Richard J. Worthington

Director — AI Solution Design & Operations

Comcast · Inwood, WV · November 2021 – present

I run two teams inside the AI Enterprise Decisioning group, part of Comcast's Enterprise Business Intelligence organization. Solution design works out how decisioning strategies get enabled across the enterprise, alongside the engineers who build them. Operations designs, configures, monitors, and optimizes those strategies once they are live. Omni-channel decisioning runs through both.

• Launched the group's first offer-decisioning program, putting the data science team's models to work on customer retention. We built the system in house, and it worked.
• When the program outgrew what we had built, my team wrote the requirements for the enterprise platform that replaced it, then partnered with the engineering team that built and maintains it. My team migrated every existing use case onto the new platform.
• Spearheaded the omni-channel feedback loop, which carries the outcome of every decision back to where the next one is made. It lets us test decisioning tactics against each other in live traffic, and it produces the labeled training data that reinforcement learning needs.
• Built our own monitoring, using large language models to read the platform's logs and watch every data feed a decision depends on. It flags deviations from normal on its own and alerts the team, so a bad input becomes an investigation in minutes instead of a financial hit discovered later. That work grew into a full analytics suite my team built and calls Insight Ops: it pulls the logs several times a day, lets us ask an agent questions about them in plain language, and generates dashboards on the fly to convey a strategy's performance or an incident's impact.
• Automated our testing of tactics. My team takes a tactic's configuration, runs customers through it in simulation, and reads the results. That confirms the tactic is built the way the business asked for it, and it surfaces opportunities to hand back — where a tactic could be tuned for a particular segment or pushed toward a specific KPI.
• In 2025 the program delivered $126M in annualized incremental benefit, measured as lift against real-time holdouts in the framework above. The figure is the program's, not mine alone.

Sr. Manager — Systems Engineering

Verizon · Dallas, TX / Ashburn, VA · August 2015 – October 2021

• Created and led Team Catalyst, the systems engineering team that built the platform and methods Verizon used to operationalize AI and machine learning models.
• With my team, built Catalyst, the change-management front end for decisioning, from intake through monitoring, across wireline and wireless.
• Led the product, design, and data science teams at different points, alongside systems engineering.
• Led the design and implementation of multiple real-time decision engines for customer-facing decisioning.
• Set the enterprise standards for deploying customer and agent experiences automatically, with the first real-time feedback loop powering predictive and adaptive models at scale.
• Ran platform strategy for A.N.N. (Artificial Neural Network) and S.O.I. (System of Insights).
• Deployment velocity rose 10.5× in wireline and 3.5× in wireless over eighteen months.
• The platform went from $47M in annualized benefit to nearly $327M.