Mixed-skill on purpose
Engineer + exec + ops + curious. Nobody is the most senior or the most junior at the table. Awkwardness has nowhere to land.
Claude workshops for organizations doing the rollout.
Three practitioners. Half-day to two-day. Up to one hundred leaders shipping with Claude by the end of the session — including the ones who haven't opened a terminal in a decade.
In the room
AIE World's Fair · SF · Mastra workshop
3crew
Practitioners on-site
100leaders
Per collaboration-week session
½–2days
Shape to your week
15–40skills
Shipped live, in your repo
§ 01
You don't have an AI problem. You have an adoption problem.
The license is bought. The pilots are running. Engineering has been on Claude Code for months. But the rest of the org is either making cool emojis with the chat or quietly waiting for someone else to figure it out — and your CEO is asking what's taking so long.
We've watched the same pattern at every org we've talked to this year. The bottleneck is never the model. It's the change management — the cultural shift from “AI is a thing engineers do” to “AI is how everyone here works.” Speakers don't move that. Office hours don't move that. A room of your peers shipping something together, with practitioners embedded inside the pods, does.
§ 02
How the workshop runs.
Engagement shapes →Workshop-shaped, not classroom-shaped. We split your room into pods of four to five, mixed-skill on purpose — exec next to engineer next to ops next to CS. Each pod has a designated keyboard owner; it rotates every fifteen minutes. The practitioners float, embedded — somebody is sitting next to you when you get stuck, not raising your hand on Zoom.
Hot seat rotates
You build, watch, build again. No “stand up and demo to the whole room.” The hot seat moves long enough to feel something, short enough that nobody owns the failure.
Practitioners, not speakers
We float through the pods, embedded. Roughly seven pods per practitioner at full scale, but in practice we go where the action is.
What we won't do
- Run a deck for an hour
- Single anyone out
- Make the engineers sit through a “what is an LLM” slide
- Leave you with a recording and a wave goodbye
What we always do
- Build inside your codebase, docs, and brand
- Pair exec + IC inside each pod
- Ship org-specific skills live, in your repo
- Surface your hottest internal champions
§ 03
Pre-flight. The part that makes day-of work.
The most expensive thing a workshop can do is burn the first ninety minutes on env setup. We've watched it kill rooms. So we don't.
T-2 weeks
Spec drop
Laptop image, install script, account checklist. Your IT team runs it; we're on Slack if they get stuck.
T-3 days
Dry run
We log into three random machines remotely and verify. Never goes clean on the first pass — that's the point.
T-0
Day-of
We arrive 90 min early. Tables labeled. Pods pre-assigned, mixed-skill on purpose.
T+7 days
Wrap
Recordings posted. Skills merged. Playbook delivered. Sync with your AI lead — we don't disappear.
§ 04
The arc · chat to code, inside one afternoon.
The single biggest shift for non-engineering leaders is realizing that the wall between “chat with Claude” and “Claude is building software for me” is about thirty minutes of practice. We walk every pod across it. Two-day intensives double up on Phase 03 and Phase 04 — that's where the real organizational compounding happens.
P01 · CalibrationT+0:00
Sterile pre-flight check
Laptops verified. Whisper Flow, Claude Code, and a pod-shared scratchpad warmed up. We do introductions during the install, not before.
Claude.aiClaude CodeWhisper Flow
P02 · DemoT+0:20
Voice coding — the first holy-shit moment
Whisper Flow plus Claude Code in a terminal. We dictate the same task three different ways and watch the pod realize they just typed at 180 WPM without typing. This is the moment the room shifts.
Whisper FlowClaude Code
P03 · Hands-onT+0:50
The pod ships its first real thing
Each pod picks a workflow from their own job. CS picks a triage automation. Marketing picks a competitive teardown. Finance picks a Salesforce report. We unblock, suggest, get out of the way. Working code at the end of the hour — not a demo.
Claude CodeYour stack
P04 · SkillsT+1:50
Package the work as a reusable skill
The thing your pod just built becomes a Claude skill — branded, on-policy, in your org's shared repo. Fifteen to forty of these by the time we leave, depending on engagement shape.
SkillsBrand templateYour repo
P05 · ChampionsT+2:40
Cross-pod demo & champion identification
Each pod shows their skill. We tag the people who lit up — your internal champions. They go home with a 30-day playbook for keeping the flywheel turning after we're gone.
Champion roster30-day playbook
§ 05
We're still the ones who sweat.
We've each spent fifteen years learning to type fast. Now we dictate. Nick and I still text each other when something feels strange about that — when we realize we've lost a skill we used to be proud of. We get the whiplash personally. So when your hundredth leader puts their hands on Claude Code for the first time and freezes, we know exactly what they're feeling.
Compassion isn't a posture we put on for the workshop. It's the only reason any of this works. People learn when they feel safe. They feel safe when nobody at the table is going to be ridiculed for what they don't know. The room is calibrated around that, deliberately, top to bottom.
Principle 01
Mixed-skill pods. Always.
Each pod has an exec, an engineer who's been on Claude Code for months, and a couple of curious neutrals. Nobody is the most senior or the most junior at the table.
Principle 02
The keyboard rotates.
You build, you watch, you build again. There is no “stand up and demo to the whole room.” The hot seat moves every fifteen minutes by design.
Principle 03
We've been in the seat.
Every practitioner has personally been bad at Claude in front of someone who was better at it. We remember exactly what it felt like. We lead with that memory.
§ 06
The workshop is the spark. The artifacts are the fire.
A workshop that ends when the room empties is worth roughly nothing thirty days later. Every engagement ships a kit your champions can carry forward — built during the session, signed off by you, ready to share on Monday morning. The biggest of these is the tailored playbook, written for your org by name.
A playbook tailored to your team
A bound and digital playbook written for your org by name. Covers the workflows your pods discovered, the skills they shipped, the rules of your brand and policy, and the specific way your champions should run their first thirty days of office hours. Your team's voice, your team's tools.
Org-specific Claude skills
Fifteen to forty reusable skills built by the pods during the session, branded with your brand template, scoped to your stack and your policy. Each one solves an actual problem someone in the room has.
Champion roster + spark map
By end of day we hand you a labeled list of who lit up, where they sit, and what they shipped. Plus a heatmap of which departments are warm, lukewarm, or about-to-revolt — so your next moves are aimed, not sprayed.
Recordings + searchable transcripts
Every plenary recorded, transcribed, and uploaded to a private destination of your choosing. People who couldn't attend get a chaptered, searchable archive — not a four-hour video they'll never watch.
§ 07
Three practitioners. One shipping crew.
You're not hiring a speaker. You're hiring three full-time builders at companies that use Claude in anger every day. We run workshops as the side of our job that translates what we do daily into something your team can pick up by 5pm.
Zachary Proser
Full-stack engineer, 15 years shipping distributed systems, internal tooling, and customer-facing applications that automate manual tedium away. Currently one of five on the Applied AI team at WorkOS — which powers auth for OpenAI, Cursor, and most of the labs you've heard of. Ships production Claude Code daily.
- 01
- Full-stack engineer · 15 years
- 02
- Core engineer · Cloudflare (when eng was ~100 worldwide)
- 03
- Staff DevRel · Pinecone
- 04
- Open-source developer, speaker, trainer
- 05
- AIE NY & AIE London workshop instructor
- 06
- DevSecCon 2025 keynote — watch the talk →
Nick Nisi
Staff engineer and AI practitioner working at the front of the field. EM of WorkOS's DX / AI team — owns the internal harnesses other teams build on top of. Co-instructed at AIE NY and AIE London.
- 01
- Staff engineer · AI practitioner
- 02
- EM · DX / AI · WorkOS
- 03
- Co-host · AIE NY workshop
- 04
- Co-host · AIE London workshop
- 05
- Long-time JS & tooling community voice
Nick Cannariato
Senior solutions engineer at WorkOS, with a developer-tools tour-of-duty before that: Twilio, Heroku, Zapier, GitHub. Knows enterprise rollouts from inside the room — what melts, what scales, what your IT team will actually approve.
- 01
- Senior Solutions Engineer · WorkOS
- 02
- Ex-Twilio · Heroku · Zapier · GitHub
- 03
- Embedded with rollouts at Fortune-scale orgs
- 04
- Specializes in adoption + change management
- 05
- Brings the "what about IT?" sanity check
§ 08
Engagements · three shapes, one shipping crew.
Book a scoping call →Every engagement is fixed-scope, fixed-price, held the moment you sign. No hourly. No discovery phase before you can plan a budget. You pick a shape, we hold the date, we ship.
Single-team workshop.
From $15kA focused half-day for one team that already has Claude in their hands and wants to compound. Engineering teams who want the Claude Code playbook compressed. AI working groups who want voice coding and skills under their belt by lunch.
- Who
- Up to 25 attendees
- Crew
- 1 – 2 practitioners
- Format
- Hands-on
- Hands-on, in-person or remote
- One lead practitioner + optional co-lead
- 5–10 skills shipped live, in your repo
- Tailored playbook (~12 pages)
- Recording + transcripts
Org-wide rollout workshop.
From $45kThe collaboration-week shape. Built for orgs running a top-down Claude rollout — leadership offsite, all-hands programming, “everyone in one room and let’s actually move.” This is where the change management lives.
- Who
- Up to 100 attendees · pods of 4–5
- Crew
- Full three-practitioner crew
- Format
- Hands-on
- Hands-on, on-site
- Up to 20 pods, three embedded practitioners
- 15–25 org-specific skills shipped live
- Tailored playbook (~24 pages)
- Champion roster + spark map delivered day-of
- T-2wk sterile pre-flight · T+7 wrap call
Two-day deep intensive.
From $90kTwo days, full crew, deep. For organizations who want substantively more shipped, more champions identified, more workflows transformed — not just sparked. Day one mirrors the half-day arc; day two doubles down on skills, autonomous task runners, and the org-specific patterns your champions will run with for a quarter.
- Who
- Up to 60 attendees · curated
- Crew
- Full crew · embedded both days
- Format
- Hands-on
- Two full days hands-on · same crew throughout
- 30–40 skills shipped live, in your repo
- Tailored playbook (~40 pages, printed + digital)
- Department-level workflow audits
- Champion onboarding mini-workshop on day 2
- T+30 follow-up working session included
For orgs we've already shipped with.
Once we've run a workshop together and the foundation is in place, some orgs ask us to keep showing up — running monthly champion office hours, reviewing skill PRs, sitting in on quarterly AI strategy. We're selective about retainers: we only take them after a workshop, because that's how we know we can actually move the needle for your team. Monthly, scoped to your scale, starting in the low five figures.
Mention it on the scoping call →Travel for the crew is billed at cost, separately. Venue, catering, and per-diems are on you. We can deliver remote — it's effective at smaller sizes; we'll be straight with you about what changes.
In the room · AIE London · WorkOS
AI Engineer World's Fair · New York & London · WorkOS internal Claude Day · SF
§ 09
Frequently — and reasonably — asked.
Will this embarrass executives who haven't touched a terminal in years?
No — and that's not an accident. See § 05 above. Mixed-skill pods, a rotating keyboard, no public coding, no "stand up and demo." The shape of the room makes embarrassment impossible by design, not by hope.
Can you actually handle a hundred attendees?
Yes — it's the engagement we run most often. We split the room into roughly twenty pods of 4–5 with one keyboard owner each. The three of us rotate through, embedded — roughly seven pods per practitioner, but in practice we float to where the action is.
Our laptops are locked down. Our IT team is twitchy. Now what?
We send you a laptop image, install script, and account checklist two weeks before the workshop. Your IT team runs it; we're on Slack if anything breaks. We dry-run three random machines remotely three days before. The sterile pre-flight is non-negotiable for a reason.
Our engineering team is already shipping with Claude Code. Won't they be bored?
They're the engine. We pair them inside mixed pods, where they become the local expert and walk a non-engineer across the chat → code bridge. Most engineers tell us afterward it was the most fun part of their week. If you specifically want engineering-only depth, the half-day single-team shape is the right pick.
Half-day vs. two-day — which is right for us?
If your goal is to spark adoption across the org and identify champions, the half-day is plenty. If your goal is substantive transformation — workflows actually replaced, autonomous task runners shipped, your team operating in a new mode — book the two-day intensive. The collaboration-week shape sits in between.
Do you do retainers or ongoing engagements?
Yes, but only after we've run a workshop together first. We need to have set a foundation with your team before ongoing work makes sense — otherwise we'd be guessing at where the leverage actually is. Mention it on the scoping call.
Can you do this remote-only?
Yes, with caveats. The pod mechanic translates to breakout rooms reasonably well; the embedded-practitioner mechanic doesn't translate as cleanly. We discount remote engagements meaningfully, and we'll be straight with you about what changes.
How is this different from an Anthropic-led workshop?
Anthropic teaches the product. We teach what doing the rollout inside an org actually looks like — change management, the bridge from chat to code, the skills layer that turns one person's win into an org-wide habit. We co-hosted with Anthropic in SF in Feb 2026. We're orthogonal, not competitive.
Ready when you are
Tell us when your collaboration week is.
30-minute scoping call. We send you a proposal within 48 hours. We hold the date the moment you sign — typically booked out 6–10 weeks.
What to send in the first email
- Headcount and mix (eng / non-eng / leadership ratios)
- Target date range or collaboration-week dates
- What's already been rolled out internally
- The CEO's actual phrasing of what they want