I have been on a bit of a podcast tear lately, and 99% of the time, I was interviewed by a host for 45 minutes to an hour, who then did a bit of editing and posted the podcast using the RSS generator of their choice (or never published, but that is another story).
But on one of the podcasts, I never met the host!
CallumConnects is a “micro-podcast” that comes out every single day. Only two to four minutes long it aims to provide “your daily dose of wholesome entrepreneurial inspiration.” All the guests on this podcast are focused on answering one question: “What was a hurdle you have faced as an entrepreneur, how did you handle it, and what are the key learning you got from it?”
The host, Callum Laing, has a brilliant way to record his guests. He uses a Jot form that collects all the relevant information about the guest (links, bio, head shot, etc.) and includes a “record button” at the bottom of the form for guests to submit their response right then and there!!
Brilliant! You could collect 50 hours of content in 5 minutes!
The guest information is used to generate intro and outro scripts for Callum to record, and then the recordings are spliced together into the final podcast. Guest submissions are screened, but I imagine that a personal assistant or an Upwork hire does all the screening, organizing, intro/outro scriptwriting. Then Callum batches his recording sessions and hands it back for final editing and distribution. It seems to be working for Callum; according to ListenNotes, CallumConnects is in the top 5% of all pods, and the YouTube versions of the podcast typically get 500+ views.
CallumConnects is a podcast on automatic pilot. There are lots of companies and entrepreneurs who could automate a podcast for fun and profit, just by having a form and a thoughtful question:
- What advice would you give to your 20-year-old self?
- What if you could change one thing?
- What message would you like to send to yourself in the future? (kind of like future.me, but audio).
- Or even Tim Ferris’s standard, “”What message would you put on a billboard for millions to see?”
- Or even Tim Ferris’s standard, “”What message would you put on a billboard for millions to see?”
But what if we were to add a machine-learning algorithm to the mix? There are a lot of possibilities.
First, you could have the AI do all the “human” work of producing a CalliumConnects type of podcast, including initial screening of the guest recording (train it on previously recorded podcasts of the same genre). I don’t see why a Google-like AI couldn’t also record the intros and outros and handle all the cataloging and distribution.
Or you could use an AI to augment a human host. For example, you could collect the basic data about a potential guest, scrape the web for everything that the guest has published, done, etc., to be feed into an AI that would generate questions and areas of discussion that might be unusual or unique.
And eventually, the AI could interview the guest on its own. Already AIs are writing stories at places like the AP and Wall Street Journal (See Francesco Marconi’s excellent book on AI in Journalism). An AI-generated podcast would be handy in serving niche markets. For example, a weekly podcast on the mineral futures market in central Asia. Or a podcast just for the users of your company’s product or service.
Either way, I am betting that there is a good chance that an AI-assisted or AI-generated podcast shows up in your listening queue this year!