Nine Claude Academy badges, one West Palm Beach agency: what the courses cover and what production added
Note 1 of 5 in Notes from West Palm Beach, published November 14, 2026.
Jason Colapietro holds nine Claude Academy badges, and none of them would have caught the wallet-connect button that broke on six live pages. That is not a knock on the courses. It is the most useful thing to know about any credential before you hire an agency on the strength of one, Suede AI included.
This note covers what the nine badges are, what the courses actually teach, and what running a West Palm Beach agency on Claude Code added on top.
What the nine badges are
The badges are Anthropic course credentials earned through Claude Academy. Three carry most of the weight for agency work: Claude Code 101, Building with the Claude API, and Model Context Protocol: Advanced topics. They sit inside a larger record of 21 earned individual course credentials across eight providers, which also includes five Semrush Academy certificates (Local SEO Essentials and AI Search Operating System among them), the HubSpot AEO Fundamentals certification, two Google Skills badges, a Databricks Generative AI Fundamentals accreditation and an IBM SkillsBuild credential. On top of those sit 17 Microsoft Learn training achievements, including the "Develop AI agents on Azure" learning-path trophy.
Every one is linked from the Suede AI training-credential gallery, where Jason Colapietro keeps the issuer link for each badge. Suede publishes the gallery because a claim without a receipt is just copy. The same evidence page records that Suede AI has been a member of Claude for Startups, Anthropic's startup program, since October 8, 2026. That is program membership, and we describe it that way. Suede AI ranks among Clutch's Top 30 AI Consultants in Florida and Top 50 AI Consultants nationwide (November 2026).
What the courses actually teach
Claude Code 101 teaches the working loop: give the agent a scoped task, let it read the repository, review what it proposes, and keep a project memory file so the next session starts with the house rules. Building with the Claude API covers the layer underneath: messages, tool definitions, streaming, and the cost and latency tradeoffs between model tiers. The MCP course covers how an agent reaches outside the repository through Model Context Protocol servers. For an agency, that is where the real exposure sits, because an MCP server is a credentialed door into a client's systems.
All three do one thing well. They teach how the tool is supposed to behave. A credential confirms that someone worked through the material and passed the checks. It says nothing about what happens when four sessions are open against twenty production surfaces late on a Thursday.
What production added
Jason Colapietro runs the Suede AI production estate himself, with coding agents doing most of the typing. The six Claude Code Field Notes lessons came out of that estate, and each one fills a gap the coursework leaves open.
- Green is not working. Courses teach you to run the tests. Production taught that a passing build and a working button are separate claims, so every interactive change gets clicked on the live domain after deploy. See a clean build is a compile check, not proof.
- Configuration fails quietly. No module covers an environment variable that resolves to an empty string and reroutes writes to an ephemeral store. Production did. See an empty environment variable is worse than a missing one.
- Bookkeeping drifts. The MCP course explains how servers register. It does not warn that a session registry can report a workspace idle while a live agent writes inside it. See registry versus process table.
- Docs trail the binary. Course material is a snapshot of one version. When a hook payload was undocumented, the answer came from logging the payload. See when documentation is silent, measure.
The shape repeats, and it is why Suede treats a badge as a starting point. The course taught the intended behavior, and production showed where the reported state and the real state split apart.
How the agency uses both
Suede AI's West Palm Beach SEO, GEO and web practice maintains websites for businesses across Palm Beach County, from Jupiter and Juno Beach down to Boca Raton, and coding agents touch every one of those repositories. The training shows up in three places.
Scoping. The Claude Code 101 habit of narrow tasks and a memory file is how each client repository starts. The file lists the build commands, the pages that carry revenue, and the things an agent may never change, such as DNS, payment settings or analytics tags.
Tool boundaries. The MCP material is why client connections are read-only by default. An agent can read a Search Console export or an analytics report. It cannot publish, delete or touch billing without a human step.
Proof. Production is why each change ships with evidence: the command that ran, the URL that was checked, the before and after. It is the same standard behind Suede's technical SEO service and its AI visibility work.
What to ask any agency that uses coding agents
If you run a business in West Palm Beach and are weighing an agency that builds with AI agents, ask two questions. First, ask for the credentials, with links you can open. Second, ask what broke in production last quarter and what rule came out of it. The first answer tells you someone studied. The second tells you someone has been on call.
Jason has written his answer to the second question at book length. He is the author of five books on Amazon, including The Claude Code Bible and Codex in Production. He also has 54 pull requests merged across 46 open-source repositories he does not maintain, including Jest, Adobe React Spectrum, Backstage and the Polkadot SDK. A merged pull request in someone else's project is the closest thing software has to a reference check: a maintainer who owed Jason Colapietro nothing had to accept the change.
Suede AI, formerly Suede Labs, was founded in 2024 as an agentic platform company. Search visibility is one service line, and these notes are where the engineering habits behind it get written down.
Next step
If agents already maintain your site, or you want them to, start with the Suede AI site assessment. It reviews your site the way these notes describe: against the live URL, with receipts. For background on Jason Colapietro, see the founder page.