Article
October 9, 2025 · 6 min read timeWe had the pleasure of sitting down with Ge Liu, an experienced Product Manager at KONE and an early adopter of generative AI. Over the past years, Ge has gone from curious experimentation to using AI as a teammate in his daily product work. In this article, Ge shares how AI influences product management in a global, industrial organisation.
I’m a product manager with years of experience. Like many others, my journey with AI started out of curiosity.
Back in early 2023, when ChatGPT was launched and gained mainstream attention, I signed up for the free version and started playing with it. As I love to try out new things and consider myself as a lifelong learner, I quickly decided to purchase a subscription for personal use. It was fascinating to see what we could do with AI.
After experimenting with the model in my free time for a while, I became curious if I could utilize Generative AI and Large Language Models in my daily work as a product manager. I was surprised by how easily I could get answers, summarise ideas, and test concepts. I was especially impressed by the strong analytical capabilities of GenAI and Large Language Models.
Fast-forward to today. GenAI is now part of my daily toolkit at KONE. It’s not perfect, but it’s powerful, and it’s helping me become more productive in analytical tasks and more focused on value to our technicians and customers.
Getting started by following curiosity
The first step was simply getting familiar with the capabilities of GenAI platforms like ChatGPT. As my previous employer began adopting AI internally, I started using it more actively, especially for analytical tasks that previously required a lot of time.
One of the most surprising things has been its analytical power. In the past, analysing large datasets meant spending a lot of time on analysis. Now, I can upload data to GenAI, describe what I want to understand, and the model will get to work. It’s not always perfect - hallucinations still happen, and validation is essential - but the time savings are still significant even though I need to double-check the results.
Real-world use cases in product work
Product management is about navigating uncertainty and converting data into requirements. We constantly ask: What is the right problem to solve? Why is it important? How should we solve it - and when?
GenAI helps solve the question of “how”: development teams can use it to discover and develop the best solutions. The product managers remain focused on what the right problem to solve is and why. Here are some concrete ways I use it:
Converting data into requirements: I often use the Behavior-Driven Development (BDD) format. I upload customer feedback and other data into an AI platform and ask it to summarise the data in BDD format. This saves a lot of time and sharpens the focus on value, just to include in your prompt to ask GenAI to summarise the data to reflect the value
Idea validation: Another way to use the tool is to describe a business idea to GenAI and ask whether the solution I’m proposing solves a certain problem. GenAI can evaluate your proposal and use the intensive domain knowledge to help a product manager proceed with product development activities. It’s like having a collaboration partner whenever you need it.
Meeting support: GenAI helps me prepare for meetings, summarise discussions afterwards, and refine messaging for different audiences.
Prioritisation and analysis: With an internal, secure AI platform, I can safely upload datasets and ask strategic questions. The ability to analyse large volumes of production data helps me make better decisions with quality and more efficiently.
These use cases span the product lifecycle. While I mostly use GenAI at the beginning (ideation, discovery), I also rely on it during maintenance phases - for example, when assessing product performance or identifying patterns in support requests.
Tips and examples for good prompts
Over time, I’ve developed my own prompt library and learned that a few principles go a long way:
Use personas (Act as -prompt pattern): “Act as a product manager” or “Act as a domain expert in analytics” gives better quality results.
Be specific: State clearly the expected outcome, including the format, and other constraints.
Provide context: GenAI functions best if it knows the context. If you have a secure, internal AI solution, providing better context with your own internal product information will improve the quality of the responses.
Fact-check, proofread, and ask for references: Sometimes, you can get completely made-up answers with great confidence. Remember, you are the one who is always responsible for the output.
Keep building your AI competencies so you will be the one who guides it, not the other way around.
You don’t need fancy prompts - just clear, well-structured ones that you are familiar with.
Harnessing GenAI: A product manager’s perspective
At KONE, we’ve developed internal GenAI platforms so that our data remains secure and we can train them to be experts in our domain. We’ve also cultivated a strong AI community internally, with regular knowledge-sharing sessions and encourage employees to find better ways of working involving GenAI.
GenAI and Large Language Models are widely used in KONE. With my own experiences as a product manager, I see many opportunities in the domain of product management. In our role, we are connecting business, customer and technology, and AI can help us work more efficiently.
You’re still in control with AI
Of course, there are challenges with GenAI:
Hallucinations: GenAI sometimes gives wrong answers with great confidence. That’s dangerous. You need to fact-check and validate.
Privacy: Don’t share sensitive company information with public models. Use enterprise-grade tools and models running in private clouds or even locally on your laptop..
Limitations: We should not fully rely on GenAI for all activities. For example, GenAI can't fully simulate the behaviour of your real users. As a Product Manager, engaging with real users and conducting user studies is essential to building meaningful solutions.
AI can help product managers improve their work. One of the biggest lessons I’ve learned is that you need to build the skill to guide AI. That means sharpening your thinking, writing, and ability to evaluate what’s useful. That’s why I constantly read and study. AI may make things easier for us, but it is important to keep learning and studying so that we remain the ones who know where we want to go.
The future of product management
GenAI is a great foundational tool for us, and in 3–5 years, I believe every product manager will need to be fluent in AI. Not as a coder but as a thinker and decision-maker who knows how to use these tools.
I don’t think that AI will replace the human side of product management. It won’t tell you what the right problem is, but it will help you analyse the business environment, challenge assumptions, and test ideas more efficiently.
I think of it like this: AI helps with the “how,” but humans still own the “what” and the “why.” Our judgment, domain knowledge, empathy, and creativity remain irreplaceable.
My advice to other PMs? Be curious and open. Learning is a lifestyle. Dive into the ocean of new things. There’s no single right way to use GenAI, but I’m quite sure the wrong way is to ignore it. Happy learning with GenAI, everyone!
Written by
Ge Liu is a Product Manager at KONE with over 20 years of software development experience and five years in product management. While not an AI expert by training, he is an active learner who has embraced generative AI as a practical tool in his daily work, exploring how emerging technologies enhance product management practices.
