Almost all of my conversations with content folks these days revolve around AI, and more specifically, around Claude.
I just got enterprise access to Claude––all of it, including Claude Code and Claude CoWork. In the few hours I’ve spent playing with it, I have been so incredibly impressed. Skills are so easy to create, and they make my outputs so much better. I’ve already created voice and tone skills for my execs, jobs to be done skills, brand voice and tone skills, and so much more.
I often get push back from everyone here when I write a newsletter only about AI, especially if it's super positive about it. But I am floored with how much work quality work I’ve been able to get done with Claude in such a short period of time.
Meanwhile, a good friend of mine in the industry has been sending me voice notes about Claude Code. He’s fully bought in, and is wary of being an “AI bro,” but it's saving him hours a day and sharpening his strategic recommendations, too.
I feel a bit late to the party on this. One friend told me her and her husband are using Claude to help with family financial planning, and then asked for my use cases. Oh, um, just writing and automation and strategy documentation, I said. The easy, baseline use cases it seems.
All of that to say, it has finally clicked for me. All of the talk about creating little agent versions of yourself to do various tasks––I can see how that will work, and how Claude can make that possible in only a few hours of organization. This episode from Lenny’s newsletter is a great example. It focuses on design/creative work, not writing, but the takeaways apply. For instance:
- Create dueling AI agents to build better code.
- Use AI to prevent dropping the ball at work.
-
Build personalized apps that solve your own problems first.
Now, one thing I’ve been doing in Claude to improve its writing outputs is to give it a story arc to adhere to. As I’m sure y’all experience, too, a lot of folks will send half baked blogs my way that were clearly written by AI. I often need to take those assets, figure out what it is we are trying to say––really––and rewrite the asset entirely so that it doesn’t read like AI, is positioned as thought leadership, and reads in a compelling way for a human.
You can’t prompt any of the AI tools with “just make this better,” or “rewrite this in so and so’s voice” or “use these brand and tone guidelines to remove any AI language” to get a quality output. That might solve a few problems for you, but the output won’t accomplish that last goal: to read in a compelling way for a human.
So, I use story arc prompts to achieve this. Kurt Vonnegut famously visualized the shape of stories. And today, you can find hundreds of story arc variations online. You can even go deep into each of them, and define things like the 12 stages of the hero’s journey, for example.
This type of detailed request to Claude has produced some of its more impressive content reorganizing and rewriting. It gets me far closer to something compelling than any other prompt I’ve attempted.
So, that’s it for this week. Claude and its skill feature is wonderful. I’m working on building out my own little community of AI agents with it. And story arc prompts are wildly helpful for getting the quality content outputs I need to move an asset into human editing confidently.
As always, I’m curious. How are you using AI? What prompts are working for you?
