How I Use AI to Find Email Ideas Without Handing Over the Writing

I use AI to find what I have already been saying to clients, not to replace the final writing of my emails. The useful material comes from actual questions, the advice I gave and the situations people are dealing with in their businesses.

I connect that material to an email strategy with opt-in funnels, scheduled sequences and current broadcasts. The distinction matters: the sequence can handle a defined follow-up, while the broadcast is where I write about what is happening now.

I collect the conversations before asking for ideas

My source folders contain transcripts from client calls, group coaching calls, internal meetings and YouTube videos, alongside earlier marketing emails. In the setup I described, call recordings came through Fathom, and the video transcripts came through a separate data connection.

The files gave me context to work with in Claude Code, Claude Cowork or Codex. Instead of asking for generic marketing ideas, I could ask what advice I had been giving that week and look at what clients were actually asking about.

The question is what I have already been explaining

A specific client problem can make a useful email when I have a precise answer or an anecdote from solving something similar. If the person accepts the explanation, and especially if applying it helps them, there is something concrete to write about.

That is why I ask the AI to retrieve and organize my own material. The questions come from people in my target market because they are my clients. I am not starting with a blank prompt and hoping it invents something relevant.

Client Ascension related reading: the VSL breakdown shows another way I connect sales material to questions buyers actually ask.

I still write the email

I had experimented with AI-written emails and disliked how they sounded. I described getting replies from readers recognizing the AI writing too. My approach was to use the source material for ideas and then write the email myself.

That is my editorial preference and experience, rather than proof that every AI-assisted draft must perform poorly. The practical distinction is between help finding the substance and outsourcing the voice of the final message.

The opt-in and the next sequence do different jobs

I described several entry points: an application funnel, free training, a lead magnet and a low-ticket checkout. The page collects contact details before the visitor reaches the next step, so there can be follow-up even when that visit does not end in a purchase.

The sequence depends on the situation. A checkout can lead to abandoned-checkout follow-up or a post-purchase sequence. A lead magnet can have its own sequence. I did not describe one sequence length as a universal conversion formula.

Client Ascension related reading: the acquisition guide covers the broader path from a traffic source to a sales conversation.

Broadcasts are where I use what is happening now

The broadcast is the one-off email, built from current calls, repeated questions, recent content and feedback. I compared it to a continuing series: the point is that something new is happening, not that every message was written once and forgotten.

I adjust frequency to the amount of useful material and the audience’s response. Sometimes I write several pieces ahead; often I write about what happened the day before. I described mixing advice, lighter calls to action and more direct sales emails rather than making every email the same kind of pitch.

I keep attribution separate from a claim about every purchase

The dashboard example I showed used last-click email attribution. That is a narrower measure than the belief that email also influenced buyers who eventually purchased through another path.

I argued that email contributed more broadly, but that argument was not a controlled measurement of what would have happened without email. The useful lesson here is the content and follow-up process, not a promise of a particular monthly revenue number or an assumption that sending emails has no platform cost.

Adapted by Client Ascension with AI assistance from Daniel Fazio's lesson. The first-person narrative follows the source speaker’s account. Watch the original video