AI Builds: Gmail Digest

Most AI advice for lawyers stops at prompting. Ron Drescher goes further, walking through the automated newsletter-digest system he built with make.com, Gmail, and ChatGPT — one that runs itself every morning and hands him a curated legal-tech briefing before he's finished his coffee. It's a real build, roadblocks and all, and a look at what "Jetsons" actually looks like for a solo practice.

Listen/Watch On

SpotifyApple PodcastsYouTube

Show Notes

Most lawyers think about AI as a chat box. Ron argues that's already the old way of thinking — and the real leverage is in systems that run without you, remember what they're supposed to do, and hand you a finished result before you've had your coffee.

In this episode:
- Why the podcast is shifting focus from prompting skills to persistent, automated AI workflows
- The "airport test" — Ron's framework for deciding whether an AI workflow is actually worth adopting
- Building a fully automated AI news producer using make.com
- The four-step architecture: collector, bundler (text aggregator), producer (persona prompt), and formatter
- Using a Gmail label and filter to isolate newsletter content from client and personal email
- Why persona prompting ("you are the executive producer") beats a plain summarization request
- Why GPT-4.0, an older model, was still the right tool for this specific job
- Hitting and solving the "wall of text" formatting problem with a markdown-to-HTML converter
- Using Gemini as a build partner because of its familiarity with Gmail
- Real examples of what the finished daily digest surfaced, including breaking legal-tech news

We also discuss:
- Why LinkedIn, properly configured, is a strong AI news source
- Newsletter overload and why Ron started ignoring sources he used to value
- The PlayStation 3 vs. PlayStation 2 analogy for why newer AI models aren't always necessary
- Heather Gardner's role in introducing Ron to this kind of automated workflow thinking
- Where this build actually lands on the Flintstones/Simpsons/Jetsons spectrum

Key Takeaway:

The most useful AI adoption in a small firm doesn't happen in the chat window — it happens when a workflow disappears into the background and just delivers. Ron's four-step build (collector, bundler, producer, formatter) is a repeatable pattern, not a one-off trick, and it didn't require the newest, most expensive model to work.

This episode sits right in Simpsons territory: a lawyer who's dabbled with AI tools individually but hasn't yet connected them into something that runs itself. Ron's point isn't "go build this exact thing" — it's that the leap from Simpsons to Jetsons is smaller and less technical than most solo and small firm lawyers assume.

Mentioned in This Episode:
- make.com
- Gmail (labels, search filters)
- ChatGPT / GPT-4.0 (OpenAI)
- Google Gemini
- LinkedIn
- Artificial Lawyer (newsletter)
- The Rundown (newsletter)
- Legal Tech Blogs (newsletter)
- Heather Gardner
- Apple / OpenAI trade secret lawsuit (Johnny Ive hardware device)
- Meta AI-biased layoff algorithm lawsuit

To see and download the final outputs from these tools click here.

[email protected]