The Perfect Partnership: How I Work With Claude Today
How I Found the Right Balance Between Personal Creation and Technological Editing
After a long period of trial and error with various artificial intelligence tools, I've finally found the right approach for me. I've already shared stories about my first attempts and mistakes in my previous post. Today, I want to give you a genuine behind-the-scenes look at exactly how my work with Claude functions.
When I look at what I wrote in previous stories, I realize it's no longer relevant to my writing process today. I don't use MAX or JEP anymore, though they were an important part of my trial-and-error process. Currently, I work exclusively with Claude, and I discovered something fundamental along the way.
The main problem I identified early on was that I gave Claude too much freedom in choosing words. He took my drafts and replaced simple words with "prettier" ones. Yesterday I sat down and read some of the first posts I wrote with his help, and suddenly felt something wasn't right - I had lost some of my authenticity along the way.
I read these words and don't feel they're mine. Yes, I created the original draft, but the system replaced my words with others. At the time, I thought it actually improved the text, but now I understand these words don't sound like me at all. It's like wearing someone else's clothes - perhaps beautiful, but they don't feel right.
I'm actually glad I see this now - it's a clear sign that I've matured as a creator and can better identify what's right for me and what isn't. Maybe that's the whole point: I found my true voice and can recognize it. And when I encounter something that doesn't sound like me, a red flag goes up.
This is how Claude and I work today. It all started with creating unique projects for each purpose - one for editing daily posts, another for unified posts (like my previous post about my journey with AI), another for creating the weekly newsletter, and a few other specific projects. In all these projects, I reached a point where Claude doesn't create anything from scratch. He only polishes or rearranges my content.
My work process always begins with me - and this is something I want to emphasize. I experience things firsthand, write about them in a scattered, flowing manner. The words simply pour out of me like a river, without too much thought about structure or order. Then I send Claude this draft, and he helps me improve the narrative and connections between sections.
Sometimes he adds an opening sentence, because from my perspective I would simply throw you into the middle of the situation and expect you to figure it out. "Here's what happened, figure out for yourselves how I got there." But the editor likes order, and he's more experienced than I am in these matters. There are days when I listen to him and adopt his suggestions. There are days when I simply delete everything he suggests and say, "No, this doesn't sound like me."
One of the most significant things I did was create a "forbidden words repository" - literally a long list of expressions that Claude is not allowed to use in my texts. Yes, this is completely real. This list keeps growing as I identify more and more expressions that simply don't suit me. Despite the list, Claude still occasionally tries to sneak in expressions like "journey" or "self-discovery," and I have to remind him again and again that I don't talk like that.
I choose my words carefully. I take time to go through each expression, see what I don't like, and replace it with something that feels more right. It's a process that takes time, but over time Claude also learns what I like and what I don't.
That said, I have to admit that Claude shortens many annoying processes for me. You probably know the feeling when you have a word on the tip of your tongue and you just can't retrieve it. This happens to me all the time (hey attention deficit). In such cases, I simply turn to Claude and say, "I want to say something like..." and roughly describe what I'm trying to say. He suggests several options, and suddenly I see the expression I was looking for: "Yes! Exactly that!"
Claude does much more than just improve my words, and he really helps me with everything surrounding the post. It took me a long time to refine my instructions and explain exactly what I want him to do, but now he's like a well-oiled machine. Simply a pleasure.
When I finish writing and editing a post, and I tell him the text is ready for publication, the magic begins. Within minutes, he creates a short, accurate headline, an intriguing subheading, a list of relevant tags, a meta description for SEO that helps Google understand what the post is about, pieces for publication on Facebook, Instagram, and YouTube if needed, and even a short, clear SLUG for the website address. All according to templates we created together and precise instructions (and yes, he created these instructions himself after we discussed what I wanted). And the most amazing thing - everything is created in two languages - Hebrew and English.
This is exactly what allows me to publish posts at the rate I do. I have no idea how I could do all this alone. Actually, I do have an idea - I would focus only on writing, and I would miss all the other important parts. And this is exactly why I feel that artificial intelligence has allowed me to fulfill my digital nomad dream. Because it allows me to write and live, and still produce content that works professionally.
For each type of content, I defined a slightly different work process. For instance, when I prepare the weekly newsletter, Claude already knows exactly what to do when I send him the collection of stories from the week. He reads them and prepares short descriptive sentences in the past tense, like "I told you about the hidden coffee spot I discovered" or "I took you to the strange festival I stumbled upon by accident." I also send him texts for the opening paragraph, for the "Behind the Scenes" section, and for other parts of the newsletter.
I even send him information about stories I plan to publish in the coming week, and he creates the "What's waiting for us next week?" paragraph. It doesn't always come out perfect the first time. I correct, refine, explain what works and what doesn't, and he learns from it. I edit him, he edits me, and together we reach the result that I feel is most right.
With unified posts, like the previous post about my journey with AI, the work process is slightly different. Claude has all the precise instructions, and he knows exactly what to do when I send him a collection of stories. He compiles them into one post without changing the original content, only adding opening and closing paragraphs that connect the stories and help readers understand what the text is about. We usually discuss together the length of the text, the order of presenting the stories, the balance between the different parts. But the content itself remains as it was, and that's what matters.
This process started with a lot of work. We had to get to know each other, understand how we work together. But now, after months of working together, everything flows naturally. It gets easier for me week by week, and that's exactly how it should be, right?
I am the creator, Claude is the content manager. We are a team, and as I like to think - we are really good at it. Sometimes when I look at the finished posts, I'm proud of our collaborative work. They have my voice, my experiences, my stories - but with the professional and polished design that Claude brings to the table.
If you're also writers wondering how to work with artificial intelligence in a way that doesn't harm your personal voice, here's an important recommendation: Don't let AI replace your words with other words. I learned this the hard way. Let it help you organize and edit, but jealously guard your style and personal voice. At the end of the day, that's what makes your content unique. That's what makes readers truly connect with you. And it's the only way to ensure that when you read the text a year from now, you'll feel that it was truly you who wrote it.
Your support helps me keep experiencing new worlds – and sharing them with you through my words. Sometimes, one cup of coffee can make all the difference.
Thank you for being part of this journey.
Who Am I? | The Journey Through Time | The Daily Journey
Thanks Maya. I have been thinking a lot about exactly this topic as I am exploring both Substack for writing, and AI as a coach/collaborator in many different parts of life. The timing of your tips is perfect. Even better was the account of cliff diving in the Philippines. I am so much looking forward to combining travel with AI with purpose and writing. Here's to seeing what is out there on the road!
I'm actually glad I see this now - it's a clear sign that I've matured as a creator and can better identify what's right for me and what isn't. I like it