I imagine and build practical AI solutions.

Tools and strategies which play nice with complex processes, enterprise systems, and established teams.

I also ask questions, like “is this something software can fix?” and “do we actually need generative AI for this?

An experiment:

The Woof Book

A curated encyclopedia of dog breeds.

Or: A design and scaling experiment involving tens of thousands of generated examples. Built with MidJourney, ChatGPT, and some prototype tools.

Automate expensive, joyless work.

High-value, complex, repetitive tasks tend to fall to overqualified staff.

The work can take weeks and crushes morale.

It doesn’t have to. Some examples:

Title tags and meta descriptions for an entire website in minutes.

Persuasive writing, page-relevant keywords, and SEO best-practices at 1,000× speed. Save your best people for better things.

Sales pages, at scale, in minutes.

Generates an elevator pitch, key selling points, a detailed summary, and 3 FAQs for each item in a given list of products.

Export a list of existing title tags and meta descriptions for every page of a website.

Techie corner.

The tools above use outputs from one process as inputs to another, transforming them according to my whims.

SEO Meta Scribe currently exports results to CSV. With minor tweaks it could instead publish to a live site or integrate with a fruity piece of legacy software.

This, I think, is what’s needed for generative AI to be more than just a shiny toy in mature organisations.

If you’ve thought hard about these problems, I’d love to talk.