Let AI review your product flows live👩🏻💻
Use Google's AI tool as your live assistant for visual tasks like onboarding UX challenges, or a prototype review.
Thank you for reading The Product Courier.
We’re now 2,275, which is 149 more than last month! 🙏🏻
If you're new here, welcome!
I’m Toni, I help B2B companies build scalable growth engines to deliver results fast.
👉🏻 Let’s connect on Linkedin.
With Lucas, Product Marketing Freelancer, we share actionable tips to help you deliver better work and accelerate your career by leveraging A.I. and other tools.
In this 38th edition, I will share :
The new AI assistant crushing ChatGPT - and it’s free! 💰
How you can use AI to assist you with on-screen flows or recordings. 🤯
Happy reading!
News of the week. 📰
The Chinese whale that’s taking on ChatGPT. 🐳
A recent (2023) AI company from China just released its new AI model Deepseek V3, and it’s making waves!
This new Open-Source model scored higher than Meta’s Llama 3, and matches the performances of ChatGPT 4o, Claude Sonnet 3.5, and OpenAI o1 (complex reasoning model).
It includes all the latest AI features recently released by the big players.
It even outranks them in certain coding and complex reasoning benchmarks.
And it was built in record time, with much less resources than OpenAI or Meta.
Why it matters
“Ok, Deepseek is yet another LLM model available, so what?”
This is where it gets interesting. Deepseek is currently accessible for free.
Without spending a dime, you can:
run complex math/code reasoning prompts- like you would with OpenAI 1o.
do online searches referencing multiple sources.
use multimodal queries (upload docs and images for analysis).
And, if you want to build AI features, Deepseek’s API is 30x cheaper than GPT-4o! 😱
What’s the catch?
Deepseek is cool for day-to-day queries if you don’t want to pay for an AI Assistant.
But there are limitations and constraints you must keep in mind :
Deepseek is submitted to China’s censorship:
don’t bother asking any politically sensitive questions.
you cannot opt out of data training:
all conversations and data will be used to train the model.
🚨 Never share any confidential, sensitive information.the Chinese government can access all the data
all your conversations could potentially be reviewed.
It sometimes thinks it’s ChatGPT.
This raises questions about the training sets used.
Nevertheless, Deepseek V3 is a great way for PMs to practice prompt writing, test AI reasoning, or run online searches with AI.
Try it out if you want to improve your AI skills for free.
Do you want to train your team to master Deepseek and other AI tools?
Check out our unique AI training offer for Product teams.
Actionable modules and practical exercises tailored for Product tasks.
Leverage AI and see instant results to increase productivity.
Book a call to define your tailored training plan and empower your team to master AI.
Our Community Poll. 📊
Your weekly tutorial. 🧰
Use AI as your live product assistant. 👩🏻💻
"Design is not just what it looks like and feels like.
Design is how it works."
Steve Jobs
✅ PMs struggle to validate UI/UX ideas quickly without dedicated expertise.
✅ Teams can get stuck with design bottlenecks, slow iterations, and cognitive biases.
✅ You can now share your screen with AI in real-time, record videos, and interact.
✅ Get instant feedback on flows and designs, evaluate demos, and more.
UX bottlenecks & biases hold you back.
Let’s be real—if you’re a Product Manager, you’ve hit this wall before.
You’ve got a cool idea or a new feature to launch, but your UI&UX team is swamped. Or there is no team because you work for an early-stage startup.
You’re left alone with tools like Hotjar, Figma, and Miro, to analyze user flows and feedback.
Connecting those user insights to actual improvements without the expertise?
That’s another story.
And when you decide to tackle those challenges and take action, you also have to battle your cognitive biases like:
Confirmation bias: You see what you want to see.
Recency bias: The latest feedback feels more important.
Selection bias: You gravitate toward feedback from familiar users.
False consensus bias: You assume users think like you.
All of this slows down iterations, makes you miss key insights, and leaves you relying on your tech team's decisions.
Time is of the essence. Waiting is not an option
Imagine, you’re working on a new onboarding flow. You’ve gathered feedback from Hotjar, and customer interviews, and have a prototype. But you’re wondering:
Where are the biggest drop-offs?
Which UI elements are confusing?
Of course, you can analyze hours of recordings and meetings, and then start crafting ideas.
And if you don’t have UX experts, you can go online and search for “UX best onboarding practices” and ask around for help.
Or you can get an AI assistant to challenge your beliefs, spot patterns you missed, and provide actionable insights, whenever you need.
No more waiting. No more guesswork. Just get s**t moving.
Enter Google AI Studio
Google AI Studio is Google's web interface for developers to quickly prototype and test AI applications by:
Fine-tuning and customizing Gemini AI models
Testing prompts directly in a chat interface
Getting API keys instantly
🆕 Interacting with the Gemini models through multimodal live input.
This new feature lets you interact directly with AI by:
talking directly to Gemini (and getting text or audio responses)
showing your environment with your webcam
sharing your live screen and what you’re working on
It goes beyond uploading static interface or website screenshots and asking for details.
You can now have a “conversation” in real time about your onboarding flow, analyze your prototypes, review your posture during an interview, compare product demos…
And get insights from an AI assistant whenever you need them.
Let’s see in action with 3 real-world examples:
onboarding flow review
prototype UI/UX suggestions
Hotjar recordings analysis