“I need someone to quickly look up the current research on learning transfer.”
“And then a catchy text for the presentation slide, please.”
“Oh, and a suitable image wouldn’t be bad either.”
Sounds like a typical Monday morning in an L&D project team? Or just wishful thinking?
Since the new @mention feature in ChatGPT was introduced, both are possible: a daily reality without team meetings, tool switching, or the copy-paste carousel.
What exactly is the @mention feature?
Imagine having a conversation with ChatGPT – and simply bringing in a second, third, or fourth AI using @.
Each with a clear role.
Each steps in right within the ongoing context.
And: They all know the previous conversation.
Simply type @MarketingGPT, @ResearchGPT, or @DesignGPT – and the respective specialized GPT takes over its part.
📌 Important: This only works with Custom GPTs – your own specialized AI assistants that you create once and can use again and again.
Why this is so helpful – especially in a learning context
Most Learning & Development projects thrive on a clear understanding of roles:
- Someone researches the content
- Someone structures it didactically
- Someone writes, designs, and thinks in terms of target groups
- Someone checks whether the whole thing also looks good
With the new feature, you can map out exactly that – but with GPTs.
A real-world example:
- You ask @ResearchGPT:
“What does current research say about blended learning in sales?” - You forward the answer directly to @MarketingGPT:
“Please use this to draft a clear text for our customer documentation.” - And then to @DesignGPT:
“Create a landscape visual for me – preferably using a metaphor.”
Result:
A mini-team of AI assistants – in the same chat, without loss of context. And you are the conductor.
How to Build Your GPT Team
You don’t need an IT degree or API knowledge. All you need is a GPT Plus subscription and a bit of imagination. Because good GPTs are born from clear roles – not from technology.
Here is an example of a small, powerful team:
@ResearchGPT
Task: Provide studies & expertise
GPT description (in the builder): “You think like a scientist.”
@StructureGPT
Task: Sort & organize content
GPT description (in the builder): “You are a fan of logic with an eye for the essentials.”
@MarketingGPT
Task: Write texts, crisp & clear
GPT description (in the builder): “You are a copywriter with a feel for target groups.”
@DesignGPT
Task: Generate images with DALL·E
GPT description (in the builder): “You think in visuals and make things look good.”
Best practices from our projects
We now use GPT teams regularly:
- to prepare training materials
- to develop personas
- when writing learning content
- to visualize learning paths
- and yes: also for brainstorming when things get stuck
What has proven effective: Clearly state what each GPT role is allowed to do – and what it isn’t.
E.g., “Keep it brief,” “Use the technical terms from the PDF,” or “Use a direct tone, but don’t be too casual.”
The GPTs then don’t work against each other, but like a good interdisciplinary team.
But beware, the team leadership remains with you
As brilliant as it all is – a small warning: GPTs do what you tell them. No more – and no less. You are and remain the strategic authority. So: No blind autopilot. Instead, smart division of labor. And that’s exactly where the appeal lies: You lead the project – your GPT team supports you.
Conclusion
The new @mention feature is not just a gimmick. It changes how we work with AI – especially in projects where many subtasks are intertwined, such as training development.
If you set up your GPTs wisely and build a flexible team, you can really pick up speed while increasing the quality of your work.
Only one thing remains important:
GPTs do not replace didactics, target group understanding, or mindset. But they help you implement your ideas faster, in a more structured way, and more creatively. And sometimes that’s half the battle.
Curious?
Then download our compact one-pager here:
