Francois' Highlights #5 - Marketing in the AI era

A positioning masterclass using deep research | AI tools got big upgrades | Creating your AI Project Assistants | The future of content marketing

If you only check one thing today = follow Kieran’s guidance and set up your AI Projects Assistants (easy). If you're into Product Marketing, focus on ⬇️

- François

A 14-min masterclass on using ChatGPT deep research to differentiate vs hyped competitors

Best practice in positioning is to identify real and vivid user pain points and show that you deeply understand and solve them.

Carly, Director of Product Marketing and Content at Bubble, masterfully used deep research to gain a differentiation edge. She shared her detailed process and results in this lightning talk that got PMMs and CMOs to lean in big time.

Context: The hype around AI-native app builders such as Lovable forced Bubble, the no-code leader, to urgently adapt its positioning. 

Hypothesis: With its visual dev platform, Bubble can exploit the limitations of AI-natives, which ace prototyping, but struggle with the last 20% of development + scaling.

Workflow: Used ChatGPT Pro’s Deep Research to ID user perceptions - and specific words - of competitors’ weaknesses on SM, review sites, and competitors’ marketing.

Her deep research tips: 

📆 Given the speed of AI players, focus your research on the last 4 weeks 

🔦 Find & exploit disconnects between competitors’ claims and actual user experiences. Integrate users’ vivid expressions of frustration in your messaging

📝Ask ChatGPT to create comparison tables, SWOTs and market maps based on actual user perceptions. User feedback matters more than what marketers write on their sites

Carly’s approach to prompting

Here are my competitive prompts that Carly used as foundation and improved. You can ask her questions here.

My selection of tips, news and workflows

🛠️ Spring brought us big AI tool upgrades 🎁 !

I don't typically cover tool news, but the last 10 days were impressive:

Stealing the show: ChatGPT 4o can now generate stunning images

It set social media ablaze when folks shared their creations and “Oh sh#t moments”. That says a lot when image generation has existed since DALL-E and Midjourney.

Impressive upgrade, with (among others):

  • accurate rendering of text in your images (finally!)

  • impressive iteration and remixing by simply chatting

  • can even integrate images within your images

💡If you get stuck, try simply asking “Why are you stuck and can’t give me what I want?” It explained why my original image or my prompts were not right and laid out 3 different options for fixing that.

Many impressive examples in their launch post, including the image below (notice the perfect text rendering, how the brand flows on the t-shirt, and the person + Bay Bridge’s reflexion in the glass!). I also love how they list their limitations at the bottom of the post - builds trust.

Their prompt was: “A wide image taken with a phone of a glass whiteboard, in a room overlooking the Bay Bridge. The field of view shows a woman writing, sporting a tshirt with a large OpenAI logo. The handwriting looks natural and a bit messy, and we see the photographer's reflection. The text reads: {inserted_text}”

» Just like we can now create websites and apps without code - with Claude Artifacts & Claude Code, Lovable or Bubble - we can now design our own logos, standout ads, next level memes, turn our customer’s pics into action figure shots (hyper-personal ABM anyone?), and do a ton more without design software. Until we get stuck in the final yards…

» Do more yourself. Demand more of your agencies (sorry Juan…), devs and designers.

This may change social media ad creation forever

⚡️Sylvain G - a guy who shows me great AI tools

They also got significant upgrades:

  • Claude now has web search capability (yes!) with real-time insights and source citations (premium only for now. Worth $20/mth).

  • My big personal productivity boosters (typing is sooo 2024) are finally available on iOS: Wispr Flow, LLM-enhanced dictation, and Granola, the local note-taker that doesn’t join your Zoom calls visibly.

  • Google Gemini launched better reasoning models (great b/c of deep constant access to your G-drive files) and impressive image editing functionality (available in AI Studio with “2.0 Flash image gen mode” + API)

⚙️ How to boost your productivity with AI assistants for your key projects.

Would you like an “all-knowing” AI assistant for your high-priority projects? Do you need to write the same report on a regular basis?

Kieran Flanagan details how he built his to track and work on his priorities. This one is well worth watching or listening to the full podcast (15-min only) for his detailed pointers.

I like how he adds templates to each project (e.g. memo template, progress status) to streamline writing the regular reports we all hate writing.

tl;dr:
You can use ChatGPT Projects, Claude Projects, or Gemini Gems
1. Create a separate project assistant for each major outcome
2. Give it all the context. People don't understand how amazing AI is at holding all the context for you.
3. Instructions: Provide clear instructions on how to work with you
4. Templates: Give the Project AI assistant templates of frequent asks

Kieran Flanagan (see full LinkedIn post)

Kieran Flanagan

💡Content marketing will never be the same

Tomasz Tunguz explains how he now consumes “webinars”. Imagine when many consume content like Tomasz does….

tl;dr:
For many webinars, I'll send a notetaker in my place
With AI, I reconfigure and choose how to receive the content
The question isn't whether your content will become someone's digital hot dog
It's whether you're providing premium ingredients or filler.

Tomasz Tunguz, GP, Theory Ventures

» The future of Content Marketing should be about creating super original and/or hyper personalized content. We'll feed that content - tailored to each segment or reader - into beautifully designed apps that we create with no code, enhanced with super creative images.

Final Words

What's clear from these experiments is that the relationship between human expertise and AI capabilities isn't fixed. Sometimes I found myself acting as a creative director, other times as a troubleshooter, and yet other times as a domain expert validating results.

Ethan Mollick, Wharton Professor, Author of One Useful Thing

Know any busy marketing execs - or founders deep in marketing - who don’t have time to keep up? Thanks for sending them this digest.

🙏 François | LinkedIn