A 20-year operator, AI-native.
The unusual combination is not "twenty years of operator detail" or "AI-native, opinionated." It is both, at the same time, in the same person. That is the thesis. The receipts are open source.
What the buyer is actually choosing
A buyer hiring me is choosing between three categories of help. The young AI-native consultant who is fast, fluent, and has not run a P&L. The senior strategist who has run P&Ls and treats AI as a tool to delegate. And me, who has done both shifts of work and built a system that fuses them.
The substrate repository is the proof that the fusion actually happened. It is not a deck. It is eight layers of files, each composing the next, every claim cited.
The arc, compressed
- Customer support, mid-2000s. Geometry Global. IBM. Learned that the first words are never the actual problem.
- Sales, late 2000s to early 2010s. Direct, then channel. Learned that one fabricated detail unravels months of trust.
- Ops + analytics, mid-2010s. Performance, attribution. Learned that whoever owns the source system owns the metric.
- Social and content, late 2010s. Earned and owned. Learned that compounding channels are the only ones worth building.
- Product marketing, 2020 onward. Wingman, Clari, JustCall, PreSkale, Fincent. Learned that internal applause is not signal.
- AI-native, now. Substrate, Commons, MessagingGPT. Building the context that AI reads, not the prompts AI runs.
How to verify before paying
Three things. Open github.com/k3sava/substrate, read PRINCIPLES.md and one pattern of your choice. Open codex.iamkesava.com, read three tier-A insights, check their brier scores. Open web.archive.org, find the wingman homepage on jul 25 and sep 10 of 2021, count the words. The pattern across the three is the thesis, demonstrated.