Hey good folks and happy Friday! It’s basically Halloween, and I have a little princess and a lollipop on my hands. Hope you are getting excited about tomorrow night! Cross your fingers it doesn’t rain here in Austin (the only time I’ll ask this of you!).

Ok, let’s go.

I write a lot here about AI content production and processes, but in many places, AI has been so enforced as part of the working model that it’s clear some folks aren’t putting nearly as much thought into what they produce. Whether that be one-pagers, strategy docs, decks, there’s a whole lot of AI slop flying around.

One thing that I’m reminding my team, and our agencies, is that while AI is a useful tool, the manual processes by which we’ve done content marketing and SEO are still incredibly valuable. For instance:

• Double checking keyword recommendations, by hand. This means taking each keyword, typing it into Google, and double checking that it doesn’t overlap with any other keyword recommendation. This includes looking at the top ranking assets, parsing out how Google is understanding the keyword (what assets are ranking, are they all similar or not, and if not, that’s a good indication Google doesn’t know search intent for this term…), seeing what similar searches folks do for that term, double checking that you don’t already have an asset ranking for that term (even if not well) by typing in your company name + keyword into Google. All of this helps hone your keyword strategy, reduces the total number of assets you typically need, and ensures you can beat the competition.

• Building briefs and outlines, by hand. I know AI can do this “well,” but I’d argue not well enough. Building briefs by hand should capture the human nuance of what you are trying to create, what it is related to, what sources to cite, etc. And building an outline by hand should show you how competitors have built an asset, and where you want to be different. Also, an outline helps you understand if this asset is worth it at all.

• Strategy reviews and editing, by hand. This is important because LLMs are watermarking assets (at least OpenAI and Anthropic). More than that though, you need a human review to make sure the final asset matches the intent of the page. Otherwise, why the heck did you create it? And how many more assets will you need to create to hit the goal? Editing with humans is crucial to catch mistakes, hallucinations, and even to build a better AI production pipeline in general.

One thing I have noticed about myself is that I use AI and my human brain to do all of these tasks. And guess what? It takes me longer than if I used only my own brain. I do think the outcome is better, though.

I prompt the LLM. I read the output. I edit it. I talk to the LLM and tell it what it got wrong and how I need it to shift. I give it links, and documents, and context to help hone it all. I talk to it about competitors, and how I’m thinking about the overall strategy and goal. I ask it to poke holes. I ask it to make recommendations. Again, I read every word it spits out, and edit the heck out of it because it says crazy stuff––often! 

For me, AI is a wonderful tool. It makes me better at my job. Faster, though? Not quite yet. Not if I’m doing it right.

A NOTE ON THIS ADVICE

You do you!

One content marketer’s best practices aren’t always right for another one, though I do try to distill out the main concepts and core practices I believe everyone can benefit from. That said, you must use good judgment when deciding whether to take advice given from folks on the internet. I am an expert, and this advice comes from my direct experience, but I am not smarter than you, and I have nothing to gain or lose because of what you do.

Thank you so much for reading. Let me know what you think by replying to this email.

Very excited to be here with y’all. 

Tracey