Where I build and test AI before I bring it to clients.

I’m developing four AI-powered products alongside my consulting practice and ZINC: DiscoverArt, a taste-graph discovery engine for art; Predictant, AI-powered prediction tools for e-commerce and Shopify merchants; GarageWire, fitment intelligence for modified vehicles; and Dispatch, the management layer for AI-powered websites. Each one is where I build and test AI before bringing what works to clients.

DiscoverArt is a taste-graph discovery engine for art. You react to a handful of works, a visual taste engine learns what moves you, and the site opens a personalized gallery drawn from the world’s great museum collections, including artists you may never have heard of. It is live at discoverart.com.

No. The AI learns your taste; it does not make the art. Every work on DiscoverArt is a real object in a real museum, with images shared through museums’ open-access programs, and nothing on the site is AI-generated.

Predictant offers AI-powered prediction solutions for e-commerce: price optimization, promotion forecasting, competitor monitoring, margin calculation, a real-time dashboard, and a ChatGPT framework for customer interactions, with Shopify Sync connecting it all to a Shopify store. It helps merchants make pricing and promotion decisions with data instead of guesswork. It is at predictant.com.

Agentic e-commerce is shopping where AI agents do more of the work: researching products, comparing options, and increasingly completing purchases on a shopper’s behalf. For merchants, it raises the importance of clean product data, competitive pricing, and clear information an AI agent can read and trust. Predictant is focused on helping merchants prepare for that shift.

GarageWire is fitment intelligence for modified vehicles. Owners park their ride in the garage and get straight answers about what fits, what rubs, and what needs work, backed by real builds running the parts they want. The promise is simple: know it fits before you buy. It is live at garagewire.com.

Dispatch sits above the coding agent, repository, and host. Every agent-made change is previewed, routed for approval by role and risk tier, attributed to whoever or whatever made it, and reversible with a one-click restore. It also monitors the site, scores it, and drafts fixes for the team to review. It is live at dispatchvault.com.

Because the fastest way to know what works is to use it. My projects are where I test new AI tools, models, and workflows on real users and real data before recommending them. The advice clients get has been put to work, not just read about, and the lessons carry straight into my services and ZINC’s execution.

Yes. Each project is an AI-native codebase, built and maintained with AI coding agents. That is the same operating model I help clients evaluate in an AI Website Transition Strategy, so I can speak to what it takes in practice: the speed, the review process, and where people still need to be in the loop.

Yes. The usual first step is a free AI and Digital Opportunity Assessment to find where AI can create the most value in your business. If the answer involves a new AI-powered product, website, or workflow, ZINC, the agency I lead, can build it with me involved from strategy through launch.