At the End of the Couch

This is Sigmund Freud's coach. Sorry, couch, bit of a slip there. It's in The Freud Museum so you can't sit on it. But look carefully and you'll see where he sat: that's it, at the end, on the green chair, in a place where his clients (patients) couldn't see him.

Imagine Fräulein Katharina reclining here in the 1920s; Freud sketching out her childhood on a note pad. Or Sergei Pankejeff, the Wolf Man, haunted by a dream of white wolves sitting in a tree. Four years' regular analysis with Freud dug up poor Sergei's  memory of barging in on his parents - who happened to be busy at the time copulating in a canine fashion. Hence the dream. To help him cope. Allegedly.

Freud's work was hugely influential; foundational to psychoanalysis and other techniques of personal emancipation and growth. And although he was wrong about some things, he remains lodged in our cultural memory like Einstein, Ghandi or Harry Styles.

Face to Face

But that green chair still bothers me. I like to see my clients when I'm coaching. We sit face to face, side by side, or at any angle allowed by the independent coffee shop. I see the subtle non-verbals - a micro expression here, an eye movement there, an intake of breath or a tilt of the head. It all means something and it's all grist to the coaching mill. But Freud ignored that.  He might as well have been on the phone. Or chatting online. Did he need to even be in the room? Which brings me to AI (artificial intelligence).


Last week I had my first coaching session with PocketConfidant. PC isn't human. PC chats with me online. PC is an Artificial Intelligence. An AI coach. Here's an excerpt from our conversation. At this point I'm telling her about times when I'm most productive:

PC: Talking about the context of your situation, where and when does it happen?

Me: So, it's when I'm truly free, I feel autonomous - which is rare - but that's when I am fully occupied and I don't feel other responsibilities.

PC: How are you impacted by this experience?

Me: I don't know. Wanting it more? Questioning where I am? Yes, it makes me hungry for freedom to work in new ways.

PC: If we dig a little deeper, tell me what do you really want?

I chatted with PC for 20 minutes. By 10 it felt like I was on line to a fellow coach. I gained a couple of insights of equal significance to those I'd normally expect with my own professional supervisor.  There were a few lumps and bumps in the conversation but overall it was extremely effective. I loved it. So I sought out PC's creator - Olivier Malafronte from PocketConfidant to find out more. Hear me chatting with Olivier about AI coaching and the fascinating questions it poses.

Whose Agenda?

I'm beginning to see why PC works. Coaching mantra #101: 'It's your client's agenda; pause your own'. Coaches listen, challenge, reflect, but they never advise. That needs discipline when you have a solution and your client is struggling (Coaching mantra #102: 'Trust the process.') PC doesn't have an agenda. She can't; she's not conscious. She's just a very clever piece of code that's learning all the time. You know she's not steering you towards her goals; she's using tough questions to help you reach yours. Every session she completes makes her better for the next time. And one day, Olivier tells me, she will speak.


If you are a coach you might now be getting twitchy about your job. What happens when PC is better than you? And cheaper. And available 24-7, 365. And 100% trustworthy (within the limits of bank-level security and above). Don't worry yet. That's not going to happen any time soon, if at all. AI experts disagree about when artificial intelligence will surpass human. Predictions start at 10-20 years (not going to happen) to never, with 100 years hence being a popular bet. And what if in 2118 this super coach gets free and decides the best way to ensure #101 (the agenda one) is to remove all agendas by destroying the planet? Again, don't loose sleep over this. One of the biggest sub branches of AI is safety research, a discipline that grapples with the practical and existential issues of keeping friendly (and unfriendly) AI on the leash.

Augment not Replace

For now PC is used to augment human coaching in three ways:

  1. To prepare for a session - the client works with PC to discover themes for their human coaching.
  2. During a session to seek clarity - the client pauses a session and uses PC to explore a difficult topic.
  3. After or between sessions  - the client checks in with PC to develop the work.

So, maybe Freud was right to stay at the end of the coach. Sorry, couch. Maybe that kept his agenda right out of his clients' faces so they could become more by talking more? Maybe a non-judgmental, trusted AI is what we need.

I wonder how far we will allow AIs to help us become more human?