The fix: abandon Canva's generator entirely, hand ChatGPT the actual reference images — Claudia's character art, John's headshot, the show's color palette — instead of describing everything in words, and let it build the scene from there. One prompt and two small revision rounds later, they've got final cover art with the title and caption built into the scene as real neon signage, no boxes, no banners. The new standing rule: gather your reference images once, then for each episode just describe the scene, and the prompt gets built from there.

From there the episode swings into two segments that have nothing to do with image generation at all — a quick one on using Wispr Flow for voice-to-text (John's doing 96 words a minute dictating versus 15 typing), and a longer one on using AI as a stand-in nutritionist: photograph a recipe's ingredient labels, hand over quantities and a serving size, get back a full calorie/fat/protein breakdown ready to drop into a health tracker. If you've ever wondered whether AI tools are actually reliable enough to build a repeatable workflow around, or you just want to watch someone hit the real walls before finding what works, this is the episode.

Key Takeaways for Podcasters Using AI

Canva's AI generator can't build a unified custom scene — no matter how detailed the prompt got, it kept pulling from stock template layouts, boxing text and artwork into separate bars instead of one cohesive image.

Canva also fought basic spec instructions — even after explicitly asking for a 3,000x3,000 square image, the generator kept returning Instagram-ratio output. Square framing turned out to be a separate lock-in step, not something the generator itself would honor on request.

Connector account-mismatches are a real, confusing failure mode — a design can generate successfully and still be invisible in your browser if the connector is authorized under a different account than the one you're logged into. The fix was disconnecting and reconnecting Canva under the right login.

Reference images beat text descriptions for likeness-matched art — handing ChatGPT actual images (character art, headshot, brand palette) produced dramatically better, more accurate results than trying to describe appearance and style in words.

The repeatable episode-art workflow going forward: gather reference images once, then for each new episode just describe the scene and idea — the prompt gets written from there and run through ChatGPT.

Even planned segments get bumped — episode 3 was billed as tackling two builds, the living continuity doc and the episode art workflow. Only the art workflow made air; the continuity doc rolls to episode 4.

AI's usefulness isn't limited to podcast production — using it as a stand-in nutritionist (photograph ingredient labels, supply quantities and serving size, get back full macro/calorie breakdowns) handles a task most people wouldn't bother doing by hand.

Persistent project setup improves output quality — a dedicated project for a recurring task, a defined role for the AI to hold, and a standing "ask if you're not sure" instruction all beat one-off prompts.

Try Wispr Flow

Try Wispr Flow: https://wisprflow.ai/r?JOHN12630 (affiliate link — I earn a commission if you buy through this)

Links

Website: https://www.podpage.com/podcasting-and-ai/

Substack: https://substack.com/@podcastingandai

Email: john.jamingo@gmail.com

This episode is honest about what last week actually cost him: a domain bought before checking if the matching social handles were even available, a Google account lockout that started with a phone-verification limit and ended with him accidentally deleting a Gmail account he actually uses, and a Claude limitation he didn't know about until it bit him — Claudia can create a Google Doc, but she can't go back and edit that same doc, which is how he ended up with seven duplicate show notes files to manually clean out. On the win side: a Drive folder from day one, Todoist wired in so action items don't get lost in chat, royalty-free intro/outro music locked down, and episode 1 actually shipped with a working RSS feed.

The episode closes with "What Else Can Claudia Do?" — John hands Claudia a real terms-of-service and privacy-policy update from an app and has her translate it into plain English: what's standard, what's actually worth paying attention to, and why two minutes of that beats blindly clicking "agree" on twenty pages nobody reads. He also breaks down the five reusable skills built for this show — copy-paste templates for all five are on Substack.

Key Takeaways for Podcasters Using AI

• Check social handles before you lock in a domain. John bought one domain, then found out he couldn't get the matching social handles — had to buy a second domain just to get those. Confirm handle availability first.
• Don't try to fix an account lockout by deleting accounts. Hit Google's phone-verification limit trying to get a Gmail for the show, deleted a few unused accounts hoping it'd help — it didn't, and he accidentally deleted a Gmail account he actually uses in the process.
• Know the tool's real limits before you lean on it. Claudia can create a Google Doc, but can't go back and edit that same one — every "update" is a new file. Plan for versioning from day one instead of finding out with seven duplicate files.
• Feed the AI your real material, not vague prompts. Paste in an actual transcript or article — vague requests get vague answers; specific ones get you something usable.
• A "skill" is a saved instruction sheet, not something to write cold. Do the task with the AI a few times first, tell it what you liked or didn't, then say "turn this into a skill" — it'll draft it and ask a few clarifying questions.
• AI can make reading terms of service actually happen. Feed it a real ToS or privacy policy update and have it flag what's standard vs. what's worth knowing — two minutes instead of twenty pages nobody reads anyway.

Try Wispr Flow: https://wisprflow.ai/r?JOHN12630 (affiliate link — I earn a commission if you buy through this)

Links

X: https://x.com/podcastingandAI

Substack: https://substack.com/@podcastingandai

Email: john.jamingo@gmail.com

In this first episode, John explains why he's making this show: most AI-for-podcasters content out there is really just hustle-culture advice about monetization and scheduling, not the actual workflow. On his other podcast, AI already cut his prep time from 8 hours to 2, and he wanted to document that process for other independent podcasters who wear every hat. So episode one is exactly that, unfiltered, including forgetting to hit record for the first ten minutes. From there, he and Claudette build the show live on mic: naming it, checking domains, locking in podcastingwithai.net, wrestling with a Google account limit that briefly cost him an account entirely, renaming an existing YouTube channel as the workaround, generating the show's branding art, and recording their first on-air trailer together.

Key takeaways for podcasters using AI:

• AI can meaningfully cut prep time; John saw his own drop from 8 hours to 2
• Check domain availability early and have a backup extension ready; .com isn't always there
• New Google accounts can hit phone-number verification limits; know this before you're mid-setup
• Renaming an existing YouTube channel is a fast workaround when a fresh Google account isn't an option
• AI image generation can build a consistent visual brand from a single reference photo
• Setting up a dedicated AI "project" space keeps tone and instructions consistent episode to episode

Links:

X: @PodcastingAndAI

YouTube: @PodcastingAndAI

Email: john.jamingo@gmail.com — send John your questions, mistakes, or ideas for the show