Selected work

Problems worth getting obsessed with.

Not every project deserves a case study. These do. Each one runs the same spine, so you can compare how I think across very different problems.

By company

Honeycomb →SoundCloud →Weather →Earlier work →

Browse

Presets

Industry

Role

Year

Showing 6 of 12 case studies

012026

AI Operations: The Evidence Changed the Question

I ran 36 discovery sessions to find where AI could help. The evidence said model capability was almost never the blocker. Context, ownership, connected systems, and trust were. So I changed the question and built the method around it.

Discovery sessions across functions

HoneycombDeveloper Tools

Read the case study

022026

GTM Operating System: Radical Frictionlessness

Isolated AI wins do not compound. I connected Claude, MCP servers, skills, Notion, Slack, notifications, and write-back into one architecture so the work moves without anyone carrying it between tools.

Connected operating architecture

HoneycombDeveloper Tools

Read the case study

032026–Present

Rebuilding Finance Close with AI

Finance spent a full week every month manually recalculating ARR and reconciling it against contract lines. I rebuilt the logic so 89% of contracts now resolve automatically, leaving only exceptions for human review.

89%

Contracts resolved automatically

HoneycombDeveloper Tools

Read the case study

042024–2026

Building AI Operations from the Work Up

I started with the work: support, trust and safety, legal, marketing, internal knowledge. Each project taught me a condition AI needed to succeed. I turned those conditions into an operating model, shared infrastructure, and a community of practice, and that became the AI Ops function.

Functions on one roadmap

SoundCloudMusic & Media

Read the case study

052023–2024

Turning Consumer Support Into a Product

Consumer support was spread across The Weather Channel, Weather Underground, and Storm Radar, across platforms, channels, and knowledge sources, measured by throughput. I rebuilt it as one product: AI entry, human support, knowledge, routing, taxonomy, analytics, and a feedback path into Product. The AI matters here, but the system is the point.

3 brands

One support product across all of them

The Weather CompanyWeather & Data

Read the case study

062018–2023

Scaling Apptegy from 800 to 4,000 Customers

Five years taking a multi-country EdTech portfolio from startup to scale, 400% market share growth and a contribution to $20M ARR, built on operating models that outlasted my tenure.

800 → 4,000

Customers

ApptegyEdTech

Read the case study

Contact

Want the longer version?

Happy to walk through any of these in detail, including the parts that did not work the first time.

Book a call →Email me →Other ways to reach me →