AI at Qualtrics
Product management internship at Qualtrics: specifically, on the AI team in the employee experience department. I wanted to learn about what product management at a big tech company could look like.
I was working on an AI-powered feature to extract actionable insights from workplace surveys. There were tens of other teams across the company wanting to build AI-powered features, and ML engineering talent became the biggest bottleneck for what could be built. When my feature ended up getting deprioritized, I still wanted to build an MVP to get in front of customers. Using ChatGPT and Claude (which only existed as chatbots at that time), I scoped and vibe-coded the most minimal version of our feature to test our question - Can LLMs can use workplace survey free-responses to write actionable insights? - and got feedback from 5 customers.
Although I was a PM intern, I did a mix of product, engineering, and design work. And since 2024, those three functions have continued to get compressed into the same IC role. Specializing in these functions was supposed to decrease the context each individual PM, engineer, or designer needed. They're getting compressed because AI is supposed to enable ICs to access more context. My gut feeling (which I've learned to trust) is that we'll only be able to meaningfully act on a subset of that context, simply because of some ceiling on meaningful effort.
There’s two kinds of context: horizontal and vertical. (Some companies might call them functional and matrix. I'm not sure why different companies have different internal lingo for the same things. We should just agree on a set of shared vocabulary.) While ICs may need more horizontal context, I think that they also need less vertical context. Instead of being a PM who touches 5 different features, they’ll be a PM + engineer + designer who owns 2. But all this is to say, "I did it first!".