Netmarketing Courses
The Magnificent 7: How I Use AI to Create Content
The Magnificent 7: How I Use AI to Create Content
Michael Carney
November 11, 2024
10:08 am
Sharing this in case someone finds it helpful.
 
Here's how I use the AI Large Language Models (LLMs) for content creation, on major projects, or for optimal brainstorming:
 
I create a brief and then post it to what I call the Magnificent 7, seven different AI LLMs: Perplexity, Claude, ChatGPT, Gemini, Meta.ai, Copilot and NotebookLM. (I cheated, and got Claude to create a Chrome extension that automatically opens browser windows to each of the AIs.)
 
Then, depending on the scope of the project (I write AI and Digital Marketing courses as well as create work for clients), I will collect all the results into a single Word doc and then create a new doc assembling what I consider to be the most useful ideas/concepts from the collective intelligence of our robot overlords. Then I add in the essential humanity by going through the composite draft, editing and modifying accordingly.
 
No, it's not fast (though, arguably, vastly faster than doing all the research myself and then constructing the content), but at least in my opinion it delivers superior results.
 
And, thanks to the multiple AIs with real-time web access, plus my own constantly-updated knowledge of what's going on in the digital marketing world, I can be confident that hallucinations are kept to a minimum.
 
Oh, and I only pay for Claude and Perplexity (because I consider them the most useful for my work), I use the free versions of the others.
 
The Process in more detail

If the project is big enough (for example, if I'm outlining a course or a course lesson), then the process I generally follow is:

  1. review all responses from the 7 AIs;
  2. paste relevant responses into a single large Word doc;
  3. save that document and then also save a copy;
  4. open both document and copy and have them sharing the screen;
  5. go to the copy and get rid of every response but the one that I believe is the best, make that the primary draft;
  6. scroll through the original, pick and choose elements from each response that add value and cut and paste those into the primary draft.

Once I've done that, I then edit, adapt and otherwise clean up the primary draft and save it as the final [unless I detect issues that need to be resolved].

Yeah, I could work with only 2-3 LLMs (and would, if I had to pay for them all). But for me the process is relatively painless (with my little Chrome extension that launches the multiple LLMs), and it's surprising how often one of the LLMs [and usually a different LLM each time] comes up with a unique idea or angle. So well worth the 30 seconds of extra effort to copy and paste the brief 7 times.

Yes, I could spend the time researching and brainstorming myself. Done that enough over far too many decades in marketing. But I finally have assistants who can do the necessary research, save me the hassle and multiply my effectiveness.

IMAGES?

Those are the tools that I use for AI-assisted writing. For images, I mostly use MidJourney (I run the same prompt 7 times, using MidJourney's Repeat feature) -- but if there's text in the image, I might use Ideogram or Flux.

AUDIO?

Mostly Elevenlabs for voiceovers. For music creation/composition/execution: Suno and Udio. For separating music tracks into vocals and instrumentals: Vocal Remover. For changing existing vocals into new voices: Kits.

VIDEO?

Too many tools to mention here, except briefly: shout out to Pictory and InVideo for AI-assisted video creation using stock footage, and to RunwayML, Lumalabs, Haiper, Kling and Minimax for AI-generated video. For lip-syncing of videos and animation of speaking avatars: StudioDiD, Hedra and a couple of others.

WANT TO KNOW MORE (AND HOW TO USE THESE TOOLS)?

Check out our AI for Marketing course.

Category : AI Content Creation AI for Marketing

Be First to Comment

Leave a Reply

Your email address will not be published. Required fields are marked *