Product Orga Fitness In The Age Of AI (EN 🇺🇸)
I hold this talk at Product Tank Munich about a topic and a question which literally keeps me awake at night: How can we use AI in our product development practice?
There have been around 100 product people attending to the recent Product Tank meet-up in Munich. I had a chance to present my practical findings and thoughts on the question:
How can we use AI tools in product development practice?
For me, it was important to provide actionable ideas and go beyond ChatGPT prompt engineering use cases like I see so-called experts focusing on this topic only.
In this article I will provide the slides including my narrative, so you can grasp everything I showed. This is what the talk was about:
Who am I?
Currently, I introduce myself as a Product Leader uncovering the value of using AI tools at work.
After 18 years of non-stop working in the digital space in different industries and different product roles and studying part-time in parallel, this year I finally took some time off.
It gave me the possibility to explore new frontiers. In my private life I started dancing West Coast Swing pretty intensive. And my professional Me started trying out different AI tools for product use cases.
This is exactly where I personally see huge potential for using the right AI tools and why I was on stage.
What is Product Orga Fitness Test?
For the talk I utilized the product orga fitness test of the Value Rebels - a boutique product consulting company on a mission to upgrade European product teams to build better digital products. This test is a quick assessment which help to understand the gaps between the status quo and the common definition of “what is a good product practice?” according to the definition of the Value Rebels.
During the talk I explored 3 of 49 questions from different dimensions with the audience using slido.com and showed how using the right AI tools can strengthen your fitness level for those topics.
Q1: Start with Problem Understanding
Here are the ratings of the live poll on the question if product teams do start with problem understanding:
So the question arises:
how could AI help you with the problem understanding?
First: why most of you did not rank with a 5? What is the reason?
There are always several reasons. But there is always this one pain, you might be familiar with: From product managers point of view this is how it looks like when you have to analyze the recordings:
After the interview you have to hear all the recordings again in full length and summarize and analyze them in order to make the insights digestible for the team and for stakeholders. So we feel that it costs us almost the whole week of work to do everything right.
As humans, we seek the excuse: I need the management to allow me to put my time in discovery. If we are honest with ourselves this excuse is one reason why we do product discovery barely.
Dovetail
Let me show you dovetail.com!
Dovetail is a customer insights hub and research platform. The free version already offers everything you need to heal the pain with making interview insights digestible really quick.
How? Let me show you.
You upload your video of the interview (it takes 1-2 minutes):
you let the tool transcribe the video (it takes 2-3 minutes):
You get suggestions on the highlight quotes of every interview (it takes 1 minute).
You go through the highlights and accept or change them (it takes 5 minutes).
If you want you can add custom tags for certain quotes on specific features, or problems, or topics.
In the end it takes 10 - 15 minutes in total per video to fully analyze and summarize the whole interview instead of watching the whole recording again. I recommend doing this work right after the customer interview, as you still remember the conversation.
When you repeat this practice for 5 videos, you are done within 1–1,5 hours to gain all insights from all interviews without the need to listen to all recordings again.
After this, you want to make your insights digestible for your colleagues:
Therefor you just let the dovetail AI create a board with clusters of quotes to different topics. You can edit them and organize them by using your own tags. Let’s say it takes you another 30 minutes to have something to work on together with your team.
And if you want to make main insights easy digestible for your colleagues, you can automatically create a summary video with all important insights from all interviews. It takes you, not longer than 10-15 minutes:
So instead of showing your team or colleagues 5 interviews 30 minutes each, you show a 10-minute video out of all significant insights, e.g., when you start with a roadmapping or planning or refinement session.
Miro AI
In case you want to do discovery work by the book, you may be want to create an interview snapshot for every interview. It’s a one-pager document introduced by Teresa Torres in her book Continuous Discovery Habits. You might be familiar with it.
I prefer and recommend using this Miro template for Interview results by David Pereira he provides for free. This template is reduced to what in my opinion is really necessary:
With this template I suggest you to take the Dovetail transcript of the interview and to use the Miro AI feature to suggest what to fill in. This is a screenshot of my very first shot with a very simple prompt:
Already with my very simple prompt I got a good enough list of opportunities I can copy, paste and edit them in the template. I got also one powerful quote which I can pretty much copy & paste. Only the facts are indeed too long for the template I use. So I either would need to re-iterate with one or two prompts, or fill out the facts by hand inspired from what Miro AI suggests me.
Automating Customer Interview Scheduling with Juttu
We saw how we can simplify the tasks of summarizing and analyzing customer interviews and make them digestible for our colleagues. But there is still another pain:
Teresa Torres suggests implementing continuous discovery habits. But scheduling all the interviews with customers and the back and forth via email to find a slot is a lot of admin effort. You could automate this as well.
Here is just an inspiration from one of a lot of YouTube videos how to automate appointment generation. You can use tools like make.com together with ChatGPT and Gmail and Calendly or another tools. There is some AI logic for text generation inside this workflow. Such automation setups are pretty common for B2B sales and customer success calls. So what I already saw working for sales or customer success, we could adapt for product discovery as well.
The downside of this solution: you need to set it up first, so it requires an initial effort.
Let me show you an alternative:
Juttu
There is a startup named Juttu.co currently in their first beta iteration out there. They solve exactly this continuous scheduling problem for you. It’s a simple mail automation with a kind of Calendly functionality and this is how it works:
You just put your list of contacts in this tool, define your interview slots in the calendar and set up the weekly limit. This tool promises to automate the continuous recruiting of customers for the interviews.
And, there is no AI inside this tool: yet, I guess.
So I want to show you, that you don’t need AI for everything in order to improve your product orga fitness level! You can start utilizing the right tools with the right methods for the proper task.
Even in the era of AI don’t forget about tools and methods without AI!
The Result: one Discovery Day, Every Week.
Let’s wrap up: this is how one day in your week could look like:
You can run 5 automatically scheduled interviews in the morning, gain insights from 5 interview recordings and create digestible summary out of them in the afternoon.
When did you manage to do it last time with 5 interviews?
Me never before!
I think: This will for sure supercharge your innovation power. And you still have 4 days a week left to do, what you’ve done before while you didn’t need to ask somebody to do discovery.
And, there is another secret, I want to tell you…
AI as Trojan Horse for Product Discovery
I see AI being the Trojan Horse for us product people to bring in product discovery to life in our companies. In my experience it’s much easier to get new tools, than to explain the whole concept of product discovery to your management or colleagues. Especially if you say: “We want to use AI” chances are higher you get support because managers consider this topic as hot and want their people to gain knowledge about it.
Q2: Time to Insight
Here are the ratings of the live poll on the question if teams try to reduce the time to insight:
Therefrom the question arises:
How could AI help to reduce time to insight?
How could AI help us to throw away bad ideas fast?
We may be want to test our assumptions and ideas with prototypes, right? But it still requires some team effort and time to create such prototypes, and it takes some days and a few people.
What if we as Product Managers could create prototypes on our own?
Let me show you Claude.ai. It’s more or less the same as ChatGPT but with a special feature called Artifacts.
Claude.ai Artifacts for Prototyping
Let assume you have an idea for a new feature or app, and you want to try out things fast. This is what I did:
I told you I started dancing West Coast Swing. And as a social dancer I think I uncovered an opportunity for a product and I want to quickly visualize my idea in order to test my assumptions. This is what I did and showed to a friend already.
I wrote down the idea of a new app and a very simple prompt. This was my first result:
It’s not bad! It’s fully clickable. I was never ever able to do such things before!
Uizard.io
Then I went to uizard.io. It’s an AI powered mockup and prototype creation tool. It is like Figma on AI steroids with a prompt command line.
In uizard.io I typed in just one sentence with the very high level idea what I wanted. Just one sentence, not the whole requirements! This is what I got within 1 Minute:
Doesn’t matter if for me or for a designer: reaching this state, overcoming the blank sheet of paper would take at least some hours or days including conversations between 2 people.
How often we just don’t create prototypes because it’s too much effort?
I think, pretty often!
Now, I could have started to iterate from here in uizard, but this is what I did instead. I screenshotted the uizard results and gave them Claude:
This is the result after less than 1 hour and may be 10 iterations:
I think it’s pretty amazing for the effort I took. I was never able to create this state of my idea visualization within such a short amount of time.
Imagine how often I can do it now!
How many different ideas I can visualize, discuss with my team or stakeholders?
Additionally, I have a full functioning react code, which I could now ask my developer to push to a server and make it available on my smartphone to run interviews and assumption tests with users.
10x Your Innovation Power
We are going to become Super Swiss Army Knifes. In a very positive way. We now can reduce dependencies within a team by augmenting our skills with the capabilities of right AI tools to iterate faster.
I am bold on it: I say, utilizing AI tools with the right methods will someday 10x your innovation power in terms of speed in uncovering business value.
And the crazy thing is: this is just the beginning. The capabilities of the technology and the possibilities of its utilization will only improve.
Imagine, what will be possible in one, two, or five years from now…
When you don’t start to decrease your cycles for customer insights and assumption tests, while others will do: you will automatically fall back in the value your product provides. And this should be a matter for both: you as a product manager in a company and you as the CEO of your own professional life.
This brings me to the next and last fitness test question.
Q3: People Development
Here are the ratings of the live poll on the question if companies do care enough about development of their people:
With these results in mind ask yourself:
How does AI affect you?
Not your team.
Not your company.
Just you and your professional role?!
Let me show you some research results…
Stanford University Study
There is a study of the Stanford University showing the following:
The blue bars are performance score of knowledge workers on tasks without using ChatGPT, the pink bars are performance score with ChatGPT.
On the left side we see results for less skilled people, on the right side we see results for more skilled people.
Now look at the green arrow:
less skilled people with ChatGPT outperform more skilled people without ChatGPT.
Look at the two red arrows:
AI as Co-Pilot allows much higher performance uplift for less skilled people than for high skilled people.
And now look at the dotted red line:
there is almost no difference in performance between skilled and unskilled people who use AI.
Having 18 years of professional experience, I count myself to the high skilled people side. So for me, it means: I need to take care about utilization of AI, otherwise fresh bachelor students will outperform me in certain tasks – not everywhere, to be fair. But still.
For young professionals on the flip side:
AI is a huge chance!
Keep this in mind, doesn’t matter which segment you belong to.
Harvard Business School Study
Another study of Harvard Business School came to the conclusion, that GPT-4 has the following impact on the productivity of knowledge workers.
The quality of work has been increased by 40 % → we can gain better results.
The speed of task completion has been increased by 25% → we can do work faster.
The task completion has been increased by 12% → we can do tasks which we have not been able to do before, like I showed you how me as product manager is now able to create prototypes.
If we put these results in a formula, we see, that productivity doubles. And this is only using GPT-4 in 2023:
Now, let’s do just a speculative thought experiment:
assume that two further technological iterations influence our productivity in the same way.
assume this happens within next 5 years, or may be 10 years: doesn’t matter. Most of us will still be working.
This means, that people, who use those tools, will be 10x productive than people who don’t.
My advice: Take it seriously! Don’t trust that companies will take your development in their hands. They don’t do it enough right now already, as we saw it in the poll results. It’s your own responsibility!
One action, you can take, is following me on LinkedIn and subscribing to my newsletter ;-) Every new follower and subscriber motivates me to keep uncovering the value of AI tools for the product development practice:
The AI Tool Radar for Web Product Development
Since I know there are a lot of tools out there, it is a challenge to follow them all and recognize when and what to use it for.
This is why I started creating such a radar of AI tools for product development, where I want continuously evaluate AI tools for our use cases in product practice.
With this radar I want to focus on bridging the gap between technological possibilities and practical business applications that prepare product teams and product people to adopt and benefit from AI tools. I want to curate suitable tools and evaluate how they can help the product community.
If you want to get the radar in high resolution and stay up-to-date with the recent versions, consider becoming a subscriber! I am working on my next post I want to publish on October 1st, with the October edition of the radar containing additional tools. I intend to keep on updating this radar over time.
Run the Full Fitness Orga Test
Behind this QR code you will find the full fitness orga test with all 49 questions. I highly recommend you going through the test for yourself and then also with your team. It helps you to start the internal discussion on where you want to improve. It doesn’t matter if with or without AI: there is always room for improvement.
How to Stay in Touch
If you got interested, this is how you can reach me and join the discussion: either on LinkedIn or subscribing to my free newsletter:
If you want to shape the future of work in product development together with me, talk to me or write me on LinkedIn or an e-mail.
If you know someone in a company who wants to start a practical exploration project to measure and learn about the impact of AI in product practice together and share the results with the community, please let me know!
Thank you for the attention, and I am open for questions and the discussion. So, feel free to comment:
Stay curious and just do it!
Alexej
