Skip to main content

AI in Product Development?

Listen:

I like product ideation brainstorming—done right and focused, it opens my mind to think much more analytically about an idea, its development, and its trajectory. But on the other hand, I often had brainstorming sessions, and they were just a waste of time. And to be honest, can you count how often a session went sideways, got stuck in the same old thought patterns, and the loudest voices in the room dominate the conversation? 

I did a test yesterday with GPT-4o, and it blew the lid off my creative potential. I had tried the same exercise with the earlier models, and it was a colossal waste of time and energy.

Adding AI To The Product Team, worth?

Short, after the test, yes, it's definitely worth. Why?

We as startup founders, product managers or developer, our job isn't just about executing on a roadmap, we have to build the roadmap and come up with the right product idea at the right time - in the first place. That means staying ahead of the curve, spotting opportunities where others see blank space. Here's how I discovered that the use of AI becomes my "unfair advantage":

  • Idea Generations: AI models are trained on massive amounts of data, so they can generate novel ideas that you might never have thought of on your own. The newest one from OpenAI is a multimodal model, which means it can use different sources as combined input, like images, text and video, at the same time.
  • Data-Backed Inspiration: Use it as a market research tool, even emulate customer feedback, or competitor analysis to spark ideas rooted in reality, not just pie-in-the-sky thinking.
  • Rapid Prototyping: Quickly draft product descriptions, taglines, or even mockups for early testing and validation. Does the idea solve the problem users and the market have? Does it actually help to make things better or it is just a technology ride?
  • Team Collaboration: In brainstorming sessions AI spark new ideas, create discussion points, and even enhance ideation across the whole team. We ran an A/B test on a new product sales strategy: We don't sell technology, what if we concentrate on customer reviews and use our technology to help?

Practical Tactics (How I see AI as CPO)

Prompting a fresh instance of any AI tools and hope for the best is not working, trust me. You need to train and explain how you think, what your current market and customer / user base tells you, how the market moves and what capabilities you and your team have. Here's some kind of thought process matrix:

  • The Problem: Give your AI model a clear prompt that outlines the problem you're trying to solve. Don't just ask for "product ideas." Get specific about your target audience, their pain points, and the desired outcome.
  • The Idea: Ask the AI to generate a list of potential solutions, feature ideas, or even new product categories. Don't be afraid to experiment with different prompts and constraints to get a wider range of results. Use direct prompts to initiate market research and sources.
  • The Prototype: Use the AI to draft product descriptions, marketing copy, or even basic wireframes. This will help you visualize and communicate your ideas more effectively.
  • The Devil's Advocate: Challenge the pros and cons of a particular idea and ask it to critically analyze it. This can help you spot potential pitfalls and refine your approach. That can be done in a role play, see my example for Apache Wayang.

How I Prompt GenAI

I start with a role play, I'm the one who challenges an idea, and the AI tries to convince the team with data driven market insights. The software in question is a federated learning stack, Apache Wayang, and we want to define one major product or use-case which decouples the project from a pure technology driven approach. I used this prompt: 

Lets's do a role play, I challenge a product idea and you are the "AI-Powered Advocate", kind of. Let's go:  

Companies struggle to manage data silos and collaborate across organizations, it's a mess and might slow AI/ML development down, in the best case. The idea now is to add a AI feature store concept to Wayang, which can transform disparate data sources into a collaborative feature store, enabling secure collaborative model training without data ever leaving its source.

Output:

GPT output screenshot for product ideation

As you see, we have numbers, sources and a role play which points to multiple decision matrix hooks. We used this output to iterate over potential scenarios for Wayang and how we build a feature store for the project. Saying this - the experiment was a success from my point of view, and we'll make use of our AI bot more regularly. 

TL;DR

GenAI is a powerful tool, but it's not a substitute for human creativity and critical thinking. Use it to supplement your own ideas, not replace them entirely. Also, be aware of potential biases in the AI's output and always vet the results carefully. I also tend to integrating AI into my XOps framework, the collaborative nature of XOps ensures that these AI-generated ideas are thoroughly vetted, refined, and seamlessly integrated into the product development process.

My Take

Leveraging AI for product ideation isn't just a "nice to have." It's a competitive advantage. If you want to build innovative, customer-centric products that solve real problems, genAI should be in your toolkit. Don't be afraid to experiment and embrace this new frontier of innovation.

Comments

Popular posts from this blog

Deal with corrupted messages in Apache Kafka

Under some strange circumstances, it can happen that a message in a Kafka topic is corrupted. This often happens when using 3rd party frameworks with Kafka. In addition, Kafka < 0.9 does not have a lock on Log.read() at the consumer read level, but does have a lock on Log.write(). This can lead to a rare race condition as described in KAKFA-2477 [1]. A likely log entry looks like this: ERROR Error processing message, stopping consumer: (kafka.tools.ConsoleConsumer$) kafka.message.InvalidMessageException: Message is corrupt (stored crc = xxxxxxxxxx, computed crc = yyyyyyyyyy Kafka-Tools Kafka stores the offset of each consumer in Zookeeper. To read the offsets, Kafka provides handy tools [2]. But you can also use zkCli.sh, at least to display the consumer and the stored offsets. First we need to find the consumer for a topic (> Kafka 0.9): bin/kafka-consumer-groups.sh --zookeeper management01:2181 --describe --group test Prior to Kafka 0.9, the only way to get this in...

Beyond Ctrl+F - Use LLM's For PDF Analysis

PDFs are everywhere, seemingly indestructible, and present in our daily lives at all thinkable and unthinkable positions. We've all got mountains of them, and even companies shouting about "digital transformation" haven't managed to escape their clutches. Now, I'm a product guy, not a document management guru. But I started thinking: if PDFs are omnipresent in our existence, why not throw some cutting-edge AI at the problem? Maybe Large Language Models (LLMs) and Retrieval Augmented Generation (RAG) could be the answer. Don't get me wrong, PDF search indexes like Solr exist, but they're basically glorified Ctrl+F. They point you to the right file, but don't actually help you understand what's in it. And sure, Microsoft Fabric's got some fancy PDF Q&A stuff, but it's a complex beast with a hefty price tag. That's why I decided to experiment with LLMs and RAG. My idea? An intelligent knowledge base built on top of our existing P...

AI's False Reality: Understanding Hallucination

Artificial Intelligence (AI) has leapfrogged to the poster child of technological innovation, on track to transform industries in a scale similar to the Industrial Revolution of the 1800s. But in this case, as cutting-edge technology, AI presents its own unique challenge, exploiting our human behavior of "love to trust", we as humans face a challenge: AI hallucinations. This phenomenon, where AI models generate outputs that are factually incorrect, misleading, or entirely fabricated, raises complex questions about the reliability and trust of AI models and larger systems. The tendency for AI to hallucinate comes from several interrelated factors. Overfitting – a condition where models become overly specialized to their training data – can lead to confident but wildly inaccurate responses when presented with novel scenarios (Guo et al., 2017). Moreover, biases embedded within datasets shape the models' understanding of the world; if these datasets are flawed or unreprese...