By skyforbes I work as a Social Media Manager / Web Marketing Assistant for a food company that specializes in sauces (mustard, hot sauces, etc.). I use ChatGPT Plus every day to help with:
- content creation (posts, format ideas, newsletter copy, etc.)
- social media strategy
- some web marketing / e-commerce tasks for several brands in the group
I’ve been trying to structure my AI usage and basically “train” ChatGPT on my ecosystem. To do that, I’m relying a lot on the Projects feature.
Here’s how I’m set up:
- I have one Project for each brand in the group
- inside each Project, I keep all chats related to that brand: social posts, campaign ideas, emails, website copy, SEO drafts, etc.
Each Project has its own brand-specific instructions, for example:
- brand voice and tone (more “terroir” / mass market vs more premium / gourmet)
- how to address the audience (formal vs informal, humor or not, level of language)
- context: positioning, key products, target audiences, main channels, etc.
On top of that, I also use global custom instructions at the account level, where I specify:
- who I am (role, main responsibilities, industry context)
- how I want the AI to respond (tone, level of detail, structure)
- that I’m managing several brands with very different identities
The idea is basically:
- Projects handle the “brand identity” and very specific context
- Global instructions handle “me” and how I work with the AI
This already feels way better than using one random chat for everything, but I’m pretty sure I can push this much further:
- save more time
- avoid repeating the same information
- better organize / version my key prompts, etc.
So I’m curious:
👉 Does anyone here use a similar workflow (Projects + global instructions)?
👉 Any tips, examples, or best practices to make this kind of setup more efficient and scalable in a multi-brand social media / web marketing context?