First of all, wow this newsletter really blew up! Thank you for subscribing! In this article I’d like to share pieces of our AI journey and what’s worked will in the hopes it will help you build and rapidly iterate on your own AI products.
AI is redefining product management by reshaping how we build, execute, and deliver products. To stay ahead, product managers need to lead, not follow. Here’s how to embrace AI and make it work for your product strategy.
AI Challenges: The Balancing Act Between Research and Delivery
AI product development isn’t like traditional development—it’s a constant balancing act between research and execution. Mastering this balance ensures you’re not just experimenting for the sake of it, but strategically identifying AI opportunities that move the needle. Avoid getting trapped in “analysis paralysis,” where you spend so much time researching that you forget to actually deliver something valuable. But don’t rush either. The dreaded “Alexa effect” happens when you release AI features prematurely, confusing users or driving them away—even if you improve the product later.
Identifying AI Opportunities: A Strategic Framework
AI opportunities aren’t just lying around—they require intentional discovery and planning. Here’s how you can get started:
Here’s how to intentionally identify where AI can drive the most impact.
1. Tune into Customer Needs and Pain Points
Start by really listening to your customers. Dive into surveys, feedback, and market research to spot recurring themes. For example, if users are having a tough time navigating your product, maybe some AI-driven personalization or a RAG (retrieval-augmented generation)-based chatbot referencing your documentation could help.
2. Check Your Data and Model Strategy
AI runs on data, but not every project requires training a new model from scratch. Consider your options: Can you leverage a pre-existing LLM, fine-tune it, or use retrieval-augmented generation (RAG) to supercharge it with your specific data? Or, could you create an agentic system that connects to company databases, BI tools, and PDFs, pulling in the right data as needed? If your data isn’t quite there yet, clean up what you have or find ways to gather more.
3. Keep an Eye on Market Trends and Competitors
For example, last week it was all about DeepSeek News, and next week, there will be another big AI development. Staying on top of these shifts can reveal market gaps and spark ideas for innovative applications. Keep up by subscribing to newsletters like mine or AI-news.
4. Align with Business Goals
Any AI initiative should mesh with your company’s big-picture objectives. Whether it’s boosting customer satisfaction, cutting costs, or driving revenue, ensure your AI plans support these goals. To get there, you’ll need buy-in from your executive team and alignment with your company-wide OKRs (if that’s the model you’re working under).
5. Mind the Ethics and Privacy
AI projects need to play by the rules. Set guidelines to ensure responsible use, tackling potential biases, and safeguarding user privacy. But hey, with all the policy changes in the US right now, it’s basically the Wild West—no rules in sight! While that sounds fun, don’t fall into the trap of experimenting recklessly. Smart AI needs smart guidelines and guardrails.
Case Study: Intercom’s AI Integration
Let’s look at a real-world example. Intercom, a customer relationship management software company, saw the potential of AI to up their customer service game. They invested $100 million into developing AI capabilities and, in March 2023, launched “Fin,” an AI customer service agent. Since then, Fin has answered 13 million questions for over 4,000 customers, including big names like Monzo and Anthropic. This move not only boosted customer satisfaction but also positioned Intercom as a leader in AI-driven customer service.
Source: Intercom supremos are right ‘in the eye of the AI storm’
Our Development Approach: Rapid Experimentation with Early Adopters
On our journey to weave AI into our products, we set up a special labs environment for our AI-enthusiast customers. This allowed us to experiment quickly and gather priceless feedback. Here’s how we rolled it out:
1. Building a Collaborative Labs Space
We invited early adopters to join our labs, giving them access to experimental AI features. This collaborative setup let us test ideas in real-world scenarios and see how users interacted firsthand.
2. Weekly Office Hours for Direct Feedback
We held weekly office hours where participants could share feedback. This regular touchpoint kept us aligned with user needs, letting us quickly adjust experiments, refine hypotheses, and scale successful ideas or scrap what wasn’t working.
3. Monthly Mini-Webinars for Education and Engagement
We organized monthly webinars to educate our early adopters on the latest experiments and upcoming features. These sessions not only informed but also empowered users to get the most out of new functionalities.
This structured approach built a community of engaged users and gave us rapid, actionable insights, speeding up our AI integration process.
From Experiments to Execution: Best Practices for Implementing AI
Once you’ve tested your ideas, it’s time to scale. Here’s how to get it right:
To successfully weave AI into your product strategy:
- Start Small: Kick off with pilot projects to test AI applications before going big.
- Collaborate Cross-Functionally: Team up with sales, marketing, data scientists, engineers, designers, and customers to continuously learn and iterate.
- Monitor and Iterate: Keep a close eye on how the product is being used and be ready to make tweaks as needed.
- Educate Your Team: Offer training so everyone understands the products capabilities, how to use it, and how to position it.
I could probably write an entire article on each of these, but for now I’ll stop here.
By thoughtfully identifying and implementing AI opportunities, we can drive innovation and deliver enhanced value to our customers, staying ahead in the competitive landscape. The product managers who experiment, learn, and adapt today will be the ones defining tomorrow’s success. Let’s keep pushing boundaries and sharing our journeys in the ever-evolving world of product management.
How have you taken an experiments-driven approach to building your products? What’s been your biggest challenge or success story with AI in product management? Let’s discuss and learn from each other!
Recent Comments