Home/Notes/AI and the human part

FROM THE HOST

AI can write the email. It can't knock on the door.

By Sondra Connor · 20 October 2026 · 4 minute read

I didn't set out to make an AI podcast. I set out to ask women what their jobs actually look like. But AI keeps wandering into the conversation anyway, like a cat that knows you're allergic. The interesting thing is that nobody I've talked to is panicking about it, and nobody is cheerleading either. They're doing what people who actually do the work always do: figuring out which parts it's good for and which parts it isn't.

In this note

  • Use it as a rubber duck, not a ghostwriter. If it does the thinking, you've outsourced the part people pay you for.
  • The human part is the whole job. It came up in almost every episode, in very different jobs.
  • AI disruption is aimed at women's jobs. Not the trades. The administrative and support roles.
  • Do more of what only works because it's you. The call, the visit, the licorice.

The licorice bucket test

Alexandria Stylianou spent a decade in automotive, including commercial fleet sales. She sold trucks by driving around Phoenix with buckets of licorice, her business card blown up and stuck on the front, knocking on doors. She bought a customer's son a remote control car at Christmas.

That sounds almost quaint in 2026. Her argument is the opposite. As AI makes sales cheaper and faster, the human part gets more valuable, not less, because on a purchase that size people still buy from people. When every inbox is full of perfectly written, perfectly forgettable emails, the person who remembers your kid's name wins.

Use it as a rubber duck, not a ghostwriter

Dr. Connie Clary teaches marketing and management, which means she's watching a whole generation of students meet AI in real time. Her test is simple: if you paste the task in and paste the answer out, the tool is doing your job. Talk the problem through with it instead. Let it poke holes. Don't hand it the pen and sign your name to whatever comes back.

That's where I'd draw the line too. If it helps you think, great. If it does the thinking, you've outsourced the one part people were paying you for.

The human part is the whole job

This came up in almost every episode, in very different jobs.

None of that is getting automated. If anything, it's what's left when everything else is.

Why AI is a women's issue

Whitney Daniels is Senior Advisor to the Governor of Arizona and leads the Governor's AI Steering Committee. She's careful about the studies, and says the right response is to sweat the data rather than the headlines. Then she gives the finding that stopped me.

The jobs most likely to be disrupted aren't the skilled trades that workforce programs are built to serve. They're administrative, clerical and support roles, and those are held disproportionately by women. Which means the standard retraining setup is pointed at a different problem than the one that's arriving.

“People can participate fully as themselves, and not having technology happen to them, but have technology work with them.”

Whitney Daniels, episode 6

Her conclusion is about who's in the room. If technology is going to be an equalizer, the people building it need to look like the people it has to live alongside.

Coming up

On 13 November, Nancy A. Shenker joins me. She's the founder and CEO of theONswitch, a former CMO of Reed Exhibitions, and she went from corporate marketing to generative AI. If you're a marketer wondering where you fit now, that's your episode.

What to actually do this week

The thing I keep coming back to

AI can write the email. You're still the one who has to show up.

— Sondra