#claude

Week fourteen, the last one: a full walkthrough of what both tools actually became — WebAI, from guided intake to a deployed site, and Image AI, from a brief to a finished visual.

Week thirteen: both tools are built, tested, and in real hands — so here's the honest list of what didn't make it, and what we would have kept going on if the clock hadn't run out.

Week twelve: the CMS milestone lands on the web side, real users push the bug count up on the image side, and a model usage log makes it clear how the tool is actually being used.

Week eleven: the tools go into other people's hands, the documentation has to teach judgment rather than steps, and the gap between done and ready gets closed.

Week ten: a CMS handoff on the web side, direct client feedback on the image side, and why the output isn't finished until the person who needs to own it actually can.

The sewing machine automated the hand sewers, not the fashion designer. Week nine on which parts of agency work AI actually replaces — and which layer becomes more important, not less.

Week eight: two tools, two builders, and what eight weeks of AI R&D inside a real agency actually looks like — less linear than expected, and more useful than a personal project could be.

Week seven: mid-term presentations put both AI tools in front of people who weren't there for the build — and the feedback exposed assumptions, not failures.

At the halfway mark, both AI builds shift from whether they work to how they ship — a look at where the image tool and Web AI stand at week six.

Week five of building two AI pipelines: why writing better prompts is a dead end, and what it looks like to build the brief into the system itself.

Week four of building AI pipelines for agency work — why execution got faster, thinking mattered more, and the real decisions happened in the doing, not the planning.

Week three of building AI pipelines for agency work — what we learned testing website tools, running brand-aware image generation, and why intake is the real problem.

A week-two look inside two AI systems we're building — a website generation pipeline and a brand-aware image system — and what they reveal about building real AI tools for agencies.

Meet Paige and Alex, two co-op students building AI systems at Goose Digital for WordPress migration and brand-aware graphic generation.

A brand new veterinary clinic, no existing digital presence, and a full marketing stack to build. Here's how AI-assisted workflows changed the timeline entirely.

From zero to a fully deployed AI showcase site in two days and about five hours of human effort — using Claude Code, Gemini, Synthesia, and Cloudflare.

How we used a Claude co-work plugin to generate 190 on-brand MCAE emails across segments and personas in a fraction of the time.

When a vendor proposed doubling prices and a 12-month lock-in across three lines of business, we built our own PDF catalog viewer in a weekend.

We built a Claude Code plugin for Cowork that generates production-ready Google Paid Search landing pages — from brief to live in about an hour.