If you only check one highlight this week: the lead story.

-François

Turning ChatGPT into a multi-player collaboration tool

Getting great results from ChatGPT on our own isn’t always easy:

1. Great prompting takes skills (even if working with reasoning models helps a ton). 

2. Giving your LLM the right context takes effort and time. Especially if you need to collect input from many sources or colleagues. 

3. Iterating is often required to get great output from your “chats” but we often run out of cycles

What if we could turn this solo sport into a team effort?

Alas, even the Team editions of ChatGPT or Claude don’t allow real multi-player collaboration.

But here’s a simple workaround I picked up from Ben Thompson (Stratechery) and Francis Brero (MadKudu)

The trick is simple: just share the chat link back and forth and build on each other’s work. 

Sounds basic—but it transforms the dynamic.

Now you’re not just one person poking at a chatbot. You’re each layering context, bringing your own expertise and ideas, tapping each other’s prompting skills, and refining outputs faster.

Here’s how it works:

1. Initial Set up and Exploration
   
   Person 1 kicks things off—gives context and goals, asks the first questions, explores a topic, runs initial research or enters templates (e.g. a QBR template, a campaign brief).
   

2. Share the Link
   
   They send the chat link to Person 2 (can be an SME, a manager, a Sales person). No extra doc needed—just the link.
   

3. Review & Audit
   
   Person 2 scrolls through the convo, seeing every prompt and response. This part’s key—they don’t just get the output, they can inspect the thinking, full prompts and context behind it.
   

4. Advance the Conversation
   
   Person 2 keeps going. Adds their knowledge, redirects the thread, or asks deeper questions.
   

5. Send it Back or Pass it on
   
   Person 2 sends the link to the updated chat to Person 1—who now has even more context to work from and maybe refined prompts and ideas.
   OR Person 2 passes it to a third person who can also add their context and or thinking.

» You’re essentially creating a live, evolving thread. Super efficient. No context lost in translation. Way better than dropping docs into Slack or back-and-forth Google comments.

This doesn’t require new tools—just a slight workflow shift.

It can be even more powerful if you initiate that chat in the context of a Project where you have attached critical background from your “marketing brain” (e.g. messaging docs, persona cards, brand writing style, etc.)

Here are five ways marketers can use this flow. Possibilities seem endless.

Preparing for QBRs - The first person, i.e. the team leader, attaches the QBR template, states the goals and the context, and adds questions for the next users to answer. She passes the link to each contributor who answers questions, adds their results, comments and learnings for their section then pass the link back to the leader who reviews and builds on that info.

Shaping Messaging with Product: Product Marketers can share a chat with a Product Manager with a messaging template, a competitive deep research and key questions to get the PM’s input on key product capabilities, differentiators and more. 

Campaign Planning - A Campaign lead creates an initial outline for a campaign, including a campaign brief with objectives, target audiences, and key messages. Shares this with other team members (PMM, Solutions Marketing) for input and adjustments. Once all have added their info, the first person asks ChatGPT to create a final brief, concepts, plans and timelines.

Content Development with an SME: A content marketer generates initial content briefs or outlines then shares with subject matter experts who add more details and pass that back to the content marketer who completes the content.

Competitive Analysis - start with deep research to analyze competitor websites and 3rd party reviews, and extract key insights. Share this analysis with a Product Manager and then a Sales leader for further context and interpretation, then hand back to PMM to compile battlecards.

Persona Development with Sales - PMM creates a buyer persona card template with ChatGPT then shares with Sales who add (via voice dictation or call recordings as attachments to go faster) extra insights and examples they find representative about that persona. Sales then share the link back with PMM who fine-tunes the persona cards.

I’m sure you’ll come up with many more. Please share your favorite :-)

My selection of tips, news and workflows

🔍 We forget: reasoning models can ace spreadsheet analysis

Inspired by Ethan Mollick’s post (see quote below), I fed one of my old P&L statements to ChatGPT o3 (as in “attached the XLS to a chat with o3 as the model”), and asked” "Give me a summary of the key insights and spot potential errors."

It “thought” for longer than I thought (3m55s), but generated:

• summary tables (quite repetitive)

• decent comments and analyses

• a list of “passed” integrity checks

• a list of a couple of errors (too small for you to report me to the IRS)

• “clean up” suggestions

• follow-up questions

Have you tried feeding it your pipeline forecast models?

That's what I'm doing next.

🛠️ I love these “Expert Critique” prompts

I'm not a graphic designer. I am not a web developer. I am not a gaming specialist. I am rarely my own target audience (well, that is actually debatable…)

So when I recently created a prototype for pokemon-style mini-app for a 9-year old with Lovable, I used the following prompts to improve my early prototype. 

The list of suggestions I got back surfaced many ideas (refactoring some code, reducing some file sizes, adding key interactions, new functionality kids love, gamifications, etc.) I wouldn’t have considered plus the associated prompts that I copied and pasted into the chat window.

Prompt:

“You are a world-class front end developer. Inspect this app in great detail and recommend the top five improvements to make this app five star out of five. Explain your rationales and provide the associated detailed prompts.

Then, do the same thing, playing the role of a:

1. world-class graphic designer, then

2. gaming UX specialist, and then

3. {whoever your target audience is}”

Final Words

Thanks for sharing these highlights with busy marketing execs around you.🙏 

Someone forwarded you this email? You can subscribe here.

François | LinkedIn 

I'm a CMO, advisor, and "CMO Wingman". Yes, that's a thing :-). Ask my clients: in this AI era, CMOs need a strategic proactive advisor more than ever. I’m former CMO at Twilio, Augment Code, Apollo GraphQL, Decibel, Udacity and Head of Marketing for LinkedIn Talent Solutions.