AI Does 90% of My Work Now as a Product Leader | Yana Welinder (Kraftful)
9 proven AI workflows to get the most out of Operator, Deep Research, and ChatGPT
Dear subscribers,
Today, I want to share a new episode with Yana Welinder.
Yana is the CEO of Kraftful, an AI product that helps 50,000 product teams extract insights from user reviews and surveys. She’s also a power user of ChatGPT and did a live demo of her top AI workflows for Deep Research and Operator.
Watch now on YouTube, Apple, and Spotify.
This episode is brought to you by Merge — Merge gives SaaS companies like Ramp and Drata a single API to launch over 200 product integrations fast. Book a meeting via www.merge.dev/peteryang and get a $50 Amazon gift card when you attend.
Yana and I talked about:
(00:00) At this point, I’m not doing anything without AI
(01:38) Live demo of Operator automating product work
(08:07) Using Deep Research for competitive analysis
(11:05) Unlocking Deep Research to get medical advice
(13:29) More everyday use cases for AI you might not know about
(19:07) The key problem with AI startups
(22:32) What should the PM actually do if AI will do everything?
(25:36) The ups and downs of being an AI founder
(29:02) How to balance prompting, RAG, fine-tuning, and evals
(32:09) The right time horizon to plan for AI
“At this point, I’m not doing anything without AI”
Welcome Yana! So I’d love to start by hearing what your top AI workflows are as a founder and product leader.
At this point, I'm not doing anything without AI.
Throughout the day, I use AI to:
Edit my social media posts
Brainstorm with ChatGPT Voice
Transcribe customer calls with Zoom AI
Shoot launch videos using AI-gen avatars from Heygen and Infinity AI
Prompt engineer with ChatGPT to improve Kraftful
The only thing I don't use AI for reviewing customer support requests, as I want to see the raw feedback to understand what users are struggling with.
Of course, I still use AI to extract the main themes afterwards.
Using Operator to automate product work
Great, so I know you’re also a power user of ChatGPT’s Operator. What does Operator do?
Operator is like giving AI your mouse and keyboard. It can book restaurants, fill out forms, and more. But I actually use it the most for this product workflow. I ask it to:
Copy all user feedback about a product (e.g., notifications) to a project
Draft a PRD and user stories based on that feedback
Create tickets in Linear based on the PRD and assign them to an engineer
So basically, the whole flow takes feedback from various sources, creates a PRD, generates a user story, converts it to a Linear ticket, and assigns it to an engineer.
Wow. What prompt did you give Operator to make this work?
Here’s my prompt:
How else do you use Operator to get product work done?
Operator is also great for:
Assessing usability. If it clicks around too much, our UX isn’t clear enough.
Running bug tests. For example, giving Operator edge cases to test.
I tweeted recently that Operator would also be perfect for watching corporate training videos.
Ha! Yes I also give it topics I’m interested in and have it scroll LinkedIn to find relevant posts. It notifies me through the browser when it finds something interesting.
So you've outsourced the doom scrolling to the computer 😅
Professional and personal use case for Deep Research
Let’s talk about Deep Research, which I think is the best AI product from OpenAI since ChatGPT. What do you use it for?
