The AI Show & Tell series โ live standalone sessions for Wix employees, timed to model releases and product news. Each one stands alone: no prior modules needed, ~35 min, and you leave with something you keep.
Every delivered session has a recording, a web deck you can page through, and its digital goodies on the same page.
Build the artifact once, open it whenever, it's never out of date. Three live builds โ including one straight off Wix's own data warehouse.
Open the session →One connection, and Claude can reach around a hundred internal Wix systems. Two live demos, and the setting that keeps it from eating your context.
Watch the session →Three doors, one Claude. Which surface for which job, how to set Cowork up properly, and what's safe to share.
Watch the session โStop re-explaining yourself. Use, build, and share Skills so Claude works the way you already do.
Watch the session โTokens, context rot, the Green/Yellow/Red context zone, resets, and the Advisor Strategy.
Watch the session โEvery session ships with digital goodies โ skills you install, prompts you paste, cards you keep open. All eight built so far, in one place.
Nothing to connected in fifteen minutes. Both install routes, the setting that stops Willow eating your context, and 30 internal tools worth your time with what each one is actually good for.
Claude can't watch a video on any surface. This gives it eyes โ what it needs in plain language, which surfaces it really runs on, and exactly what leaves your machine.
A skill that interviews you in eight questions, then writes about-me, personal-voice and working-style for you. Or take the three fill-in templates instead.
Grades what Claude still doesn't know about you โ gaps named line by line, a score per file, and ready-to-paste fixes. Re-run it any month as your files drift.
Say "run a connection pre-flight" before any automation and it checks what's actually connected. Read-only probes, pass/fail verdict, fix steps. Zero setup.
A portable prompt that interviews you about a task you repeat, then hands back a ready-to-install Skill. Works in any AI, not just Claude.
The full set of session-limit and context tricks from Session #1 on one scannable card. Keep it open in a tab.
Run it before a chat gets too long: Claude writes out everything worth carrying forward so you can start fresh without losing the thread.
Eleven sessions, ordered by build dependency โ foundations first, safety before external connections, then compositional and advanced, capstone last. Locked 2026-07-05.
Tokens and compounding, context rot, the Green/Yellow/Red zone, resetting early, and the Advisor Strategy.
Using ready-made Skills, building your own, and sharing them with a team so Claude stops needing the same briefing twice.
A decision guide for the three surfaces, a deep dive on Claude.ai and Cowork including setup, the Chrome extension as an access surface, and data safety โ what's safe to share and where the enterprise boundaries sit.
What an MCP actually is, and how Willow connects Claude to around a hundred internal Wix systems in one step. The setting that stops it eating your context, two live demos, and the checks to run before you install anything.
Turn the report you rebuild by hand into a page that re-pulls its data every time you open it. The three questions that decide whether one is worth building, three live builds including one against Wix's own warehouse, and the traps that actually bite.
The deep Claude Code session, built for people who have never opened a terminal for this.
Scheduled tasks and recurring automation, plus the agentic loop patterns that make them reliable. Requires #6.
Several Skills auto-triggering and handing work off to each other. Needs #4 and #7.
A live no-code build โ agents, subagents, skills and knowledge bases, and how the hierarchy fits together on a concrete example.
Claude as a browsing agent โ navigating, researching, extracting, filling forms. Standalone, no prerequisites.
Packaging skills, connectors and tools into one installable bundle you can hand to a team. The capstone.