I never decided to trust AI. I let it earn my trust, one small yes at a time.
I think of it as giving the smallest yes.
When I started this (AI!) business in April, I would not even connect my email to AI.
Hard no.
I didn't trust it.
Yesterday, I asked Claude to read a transcript of a client conversation (they'd granted permission for the AI recording) and drop a follow-up email into my drafts folder.
It got me 80% of the way there (to a sendable email).
Here's how I got there. (Spoiler... slowly.)
The first yes: I gave Claude the chance to help me create when I copy-pasted information in. It couldn't so much as think of looking at my inbox.
It earned my trust.
The next yes: I let it read my emails. I asked Claude to review my outbounds and draft a skill (that's "Claude" for task-specific instructions) called amy-email-voice. We went round the houses (as they say in the UK) quite a few times before the voice was right.
Finally, it earned my trust.
With subsequent yeses, I gradually let Claude have more agency. I let it write to my drafts folder. Same with transcripts. Before I let it fetch them on its own, I handed it transcripts myself and watched what it did.
Each yes was small enough that I didn't flinch. That's the whole point. The trust wasn't a decision I made up front. It was the residue of a dozen small "okay great, that worked" moments.
(Kind of like building any relationship, come to think of it.)
None of this is magic.
None of this is special.
It's just the smallest yes, granted again and again, until one day you look up and AI is drafting your follow-ups while you make tea.
So if empowering AI is a maybe someday, why not today? Pick one thing you do all day, like email, and give AI the smallest yes you feel comfortable with.
Sit with that until it feels boring. Make boring the goal.
Then, give it the next small yes.
Do it on your own, or come do it with me (find time here). Building trust step by step is exactly how I help companies gain confidence in AI, from your own inbox all the way to your customer experiences. We move through the yeses side by side.
And in my next newsletter: the Chief of Staff I set out to build in April, why it's still half done, and what I have instead that works better.