Two parallel paths to the same outcome. Mastering Claude Code for engineering teams who live in a terminal. Mastering Claude Cowork for chiefs of staff, ops leads, marketers, and execs who live in claude.ai. Same thesis — encoded judgment beats faster typing — different surface area.
The thesis is the same in both: shared context, encoded skills, specialist agents, real workflows. The medium is different. Engineers want CLAUDE.md, hooks, subagents, MCP servers. Knowledge workers want Projects, Artifacts, Skills, and a workflow they can run before the 9am stand-up. Don't make ops people learn the CLI. Don't make engineers click through a web app.
For: VP Eng, Staff Eng, Platform leads, tech leads, founding engineers. Anyone whose team ships code and reviews PRs.
CLAUDE.md as substrate. Slash commands and skills as encoded judgment. Hooks for guardrails. Subagents for review, migration, security. MCP servers wired into GitHub, your DB, your stack. End-to-end CI-integrated workflow as the capstone.
For: Chiefs of Staff, Ops leads, Marketing directors, founders, consultants, analysts. Anyone whose output is a doc, a deck, a plan, or a decision.
Everything in the claude.ai web and desktop app — no code. Projects as the team's shared brain. Artifacts as deliverables. Skills for repeatable workflows. MCP connectors for Drive, Gmail, Notion. Build a personal cowork stack that survives the next model release.
Most teams start on rung 1 or 2 and stop there. That's fine — it's the whole point of putting the courses online. But if you're rolling AI across a function or a whole org, here's the path. Each rung is designed to make the next one unnecessary, not inevitable.
Anthropic ships something material every few weeks — a model, a feature, a primitive. We update the relevant module inside two weeks of any major release and summarize the deltas in a 60-minute monthly briefing: what shipped, what it changes about the playbook, what to ignore. Free with any cohort or team license.
Reseek serves two distinct audiences with two parallel courses — and the orgs that need both. Use the cards below to figure out which door is yours.
VP Eng, CTO, Staff Eng, Platform lead at 50–2000 people. Your team lives in a terminal. PRs are the artifact. You want CLAUDE.md, hooks, subagents, MCP — and a rollout plan that doesn't rely on a champion who quits in six months. Start with the Code course →
Chief of Staff, COO, Head of Ops, Head of Marketing, founder, principal consultant. Your team lives in claude.ai, Notion, Drive, Gmail. You want Projects, Skills, Artifacts, and connectors — without forcing anyone to learn the CLI. Start with the Cowork course →
CEO, COO, Chief AI Officer with a board-level mandate. Engineering and ops both need to level up. Start with the $799 bundle so both sides share vocabulary, then run an onsite workshop ($25–35k) for each side, then graduate to a retainer if depth warrants. See the full ladder →
The difference isn't faster typing. It's whether your team's expensive judgment — code review, deal qualification, support triage, financial close — gets re-used or re-invented every time someone opens a chat window.
AI coding tools don't make individuals faster. They let a team encode and re-use the expensive judgment it already has. The companies that figure this out in 2026 will look insane to the companies that don't — the same way companies with version control looked insane to companies still emailing .zip files in 2008.
A monthly retainer designed for one team at a time — in any function. I work alongside the people who already hold the judgment, encode it into substrate and agents, and leave you with a system that keeps running after the engagement ends.
The substrate-skills-agents pattern shows up the same way no matter which team it's installed in. Three anonymized engagements:
~3 hrs/senior/week lost to PR triage. Encoded the review checklist as a specialist agent wired into CI. First-pass review went from human-bottlenecked to instant. Eleven days to ship.
Support team drowning in repeat questions. Encoded their playbook and product docs as substrate, built a draft-response agent that respected escalation rules. Median time-to-first-response cut by ~60%.
Every non-standard deal bottlenecked on one director. Encoded approval guardrails and pricing logic as an agent reps run before submission. Director now reviews edge cases, not every line.
Opinionated, self-paced, real artifacts. Below are the first nine modules of the engineering course (15 total in the paid version). The companion Claude Cowork course covers the same thesis for non-technical teams in 14 modules.
For CTOs and VP Eng deciding whether Anthropic's Managed Agents is the right call before locking in architecture. The four-way decision matrix, the seven-part runtime, three reference builds with real code (AR collections, document Q&A, fractional ops), and the production concerns most teams discover three months in.
CLAUDE.md and the first two slash commands together. Weeks 3–4: we ship them, measure adoption, and tune. By day 30 you have working artifacts merged to main and two engineers fluent enough to keep going.Reseek is Will Cousin. I work with leaders across functions — engineering, ops, support, finance, RevOps — to turn Claude and agentic tools into team capabilities, not personal hacks. My background is half-engineer, half-operator, which means I can read your codebase and your org chart in the same week.
I don't sell decks. I don't run six-month transformations. I help one team at a time build the substrate — memory, commands, skills, guardrails, specialist agents, real workflows — that makes AI tooling compound instead of evaporate. Then I install the same pattern in the next function.
If that sounds like the help your team needs, the best first step is the course above. After that, if it's worth a conversation, the call is free and direct. I'll listen, diagnose, and share two or three specific wins you could capture in 30 days. No pitch. If we're a fit, we keep talking. If not, you got a useful 30 minutes.
Two courses. Pick the one that matches the surface area your team works in. Both free to start.