Keeping up with AI tool launches is starting to feel like chasing the tide. Every day we face oncoming waves of new LLMs, APIs, and "game-changing" features — some of which are genuinely promising. But when we let ourselves get swept up in the excitement of the latest, greatest build tools, we end up with tech-stuffed Frankenproducts that don’t solve a real problem. Shiny object syndrome is alive and well.
But building great products in the AI era isn’t about grabbing every new tool that comes our way. It’s about staying focused on the horizon to discern product-market fit (PMF), and applying the right communication skills to integrate human teams with AI tools.
How to avoid the build trap
New tech has a certain gravitational pull, and gains momentum with hype from early adopters. You might start to feel that if you’re not integrating the latest model, you’re already behind. But even if you're using the most powerful AI around, it won’t matter if you haven’t nailed the problem you’re solving.
Learn from splashy AI rollouts.
Consider the momentum behind AI gadgets launched in late 2023/early 2024, especially the Humane Ai Pin or Rabbit R1. Both showcased novel interfaces and AI capabilities, only to stumble post-launch. Why? The product reviews are revealing: despite their futuristic form factor, they didn't solve existing user problems much better than a smartphone could.
These splashy products relied on a 'tech-push' approach — building something because the AI capability exists, rather than starting with a deep user need. But once the novelty wore off, their users moved on.
Now savvy product managers have an opportunity to learn from early AI product hype, and counter it with a proven method: work backward. Start with people. Map their daily struggles and hopes for tomorrow. Honestly assess: Are you spending more time browsing AI tool documentation or interviewing your customers and users? Adjust accordingly. Then, and only then, pick the tech that will actually get your intended users where they want to go. Instead of slapping the latest, flashiest LLM onto products like a badge of honor, we need to take time to find technology that speaks to the need at hand.
Actively guide team AI adoption.
I'm noticing that many product teams tend to fall into one of two approaches to AI adoptions: tech-driven or workflow-focused. Product managers have an opportunity to shape which direction their teams might take — but which is the right approach?
Let's consider a high-profile example of a tech-driven approach: the turbulent 2024 rollout of Google's AI Overviews in search results. By integrating complex generative AI directly into the core product, Google aimed to provide quick answers. But numerous instances of bizarre, inaccurate, even dangerous suggestions (like glue as a pizza topping) quickly revealed the risks of deploying powerful AI without fully nailing reliability and safety guardrails. In the rush to market, Google misjudged the core user need for verifiable information sources over potentially flawed summaries.
Now think of the approach seen in leading AI code assistants like GitHub Copilot and Claude Sonnet 3.7. Unlike generic AI chatbots aiming for mass appeal, these AI assistants target a specific developer pain point: spending too much time writing boilerplate code, searching for syntax, or implementing common patterns. To solve these problems, AI code assistants apply sophisticated LLMs trained on code directly within the developer's coding environment (IDE) to provide context-aware code suggestions and completions. While not flawless, such targeted AI applications address a concrete workflow bottleneck, demonstrably speeding up development for many users — and continue to retain users and build customer bases. This workflow-focused approach finds the right-sized solution—whether that's delightfully low-tech, or sophisticated AI used precisely.
The takeaway? AI is not your product. Solving a real problem is. Start with the why. The how will follow.
The 3 languages product managers need to speak at work now
AI makes everything messier, not simpler. Which means you need to listen and speak clearly in not one, but three key professional languages.
People
Across disciplines and departments, the product manager is the translator-in-chief. No matter what product you're working on, understanding people is key — from collecting key customer insights to mapping user interactions across workflows. To bring any product to life, you also need to communicate with data scientists, engineers, marketers, designers, and executives — all of whom speak different professional dialects. This takes practice translating complex ideas, storytelling with data, and articulating technical trade-offs. You’ll want to cover insights, technical possibilities, and strategic goals in a coherent narrative everyone can understand — and rally behind.
AI
To navigate today's AI landscape, everyone needs to get comfortable communicating both with and about AI. Functional literacy now means interacting effectively with each AI type, from LLMs and emerging agentic systems to specialized domain tools (such as AI for video generation or scientific modeling). Prompting chatbots is only the beginning — you’ll need to choose the right terms to set goals, guardrails, parameters.
The promise of AI has been widely hyped, leaving many skeptical about its practical uses. Product managers integrating AI into their work must be up front and precise about its capabilities, limitations, data needs, and risks to all stakeholders. This essential skill prevents unrealistic expectations and ensures responsible AI implementation across the board.
Brand
Increasingly, your product talks. This means your brand voice isn't just for marketing collateral — it’s expressed in every chatbot interaction, every AI-generated suggestion, every automated response. But generic AI outputs can dilute your brand. Product managers have an opportunity to champion the brand by defining AI-driven interactions to ensure they reflect the intended brand personality. When executed well, this builds trust and reinforces your brand identity. When brand voice is applied inconsistently, it creates noise and confusion, eroding user experience.
Consider the controversy around Sports Illustrated in late 2023. An investigation from Futurism alleged the publication used AI to generate articles under fake author profiles. The AI-generated content raised ethical concerns, and was also widely criticized as generic — it lacked the insight and distinct journalistic style audiences expected from the Sports Illustrated brand. The brand stumbled on two fronts: it lost trust by not being transparent about AI use, and it didn’t apply its own rigorous brand voice standards. Even legacy brands need to work hard to make their voices heard above AI-generated noise.
Refocus to rise above changing tides
Feeling overwhelmed by AI is the new normal. But don’t jump onto the next AI wave just to keep up. Instead, zoom out. Talk to users. Get curious. Listen more than you pitch. And when you speak—whether to humans or machines—make it count. Because in a world full of flashy tools, being deeply useful and clearly understood is what truly stands out. Then you won't just be struggling to ride the AI wave—you'll be the one building the surfboard.