AI as the coworker who never judges your questions
Kapila Chourey is a Product Owner at the Texas Department of Transportation, where she works in the space between business and product -- delivering digital solutions to internal teams by understanding their pain points and translating them into things worth building. Her company's larger vision, she notes with a smile, is more literal than software: a department of transportation exists to give people perfect roadways. Her own contribution is the digital layer beneath that mission, taking a problem a business unit is living with and shaping it into a solution that fits both their needs and the technical and architectural realities around it.
She has spent more than eight years in product, long enough to feel sharply how much the last two years of AI tooling have changed the job. Discovery used to mean five or six calls with a business unit just to understand what they did and where it hurt. Now meetings are recorded and transcribed, documentation is queryable, and she can walk into a new product assignment already oriented -- skipping straight past what already exists to the real conversation about what to build next. She recently picked up a new role and credits that fast retrieval of context with letting her keep up: she can query Copilot, understand the business, and arrive at meetings genuinely prepared.
Chourey has a discriminating take on the tools themselves. She reaches for Perplexity because it scrapes the whole web and shows its sources -- a source of truth that will not bend to her mood -- where she finds ChatGPT reads how she asks a question and adjusts its answers to match, agreeing too readily when she pushes back. She uses Copilot for work and Claude for anything technical, architecture or code. Pushing back on the idea that AI makes people dumb, she argues the opposite for herself: with that much knowledge in hand, she feels smarter, more curious, more able to chase down the random question -- even the ending of a confusing horror movie.
What grounds her confidence is empathy. She does not think the product owner role is going anywhere, because understanding a pain point and solutioning it well still demands a human touch and a feel for architectural limits. What genuinely worries her is everyone else -- the layoffs sweeping software, the relatives and students she knows who cannot find jobs as companies hire less and one person now does the work of three. Her resolve is not to go obsolete: she wants to keep pace with the technology and is especially drawn to learning the ethics side of AI, the part that will decide whether all this speed is pointed somewhere good.
Read full transcript of interview
And what do you do?
I work with Texas Department of Transportation as a product owner. I work very closely in between business and products. I provide digital solutions to the businesses, understand their pain points. Although my company's vision is a little different — we try to provide you guys perfect roadways, because it's Department of Transportation.
You failed this morning, but I don't blame you.
It's the rain.
Product is increasingly becoming the most interesting thing from the conversations I've been having. How long have you been working in product?
Almost more than eight years now.
You've been working in product for more than eight years. These AI tools that have come out have only kind of been a thing for the last two — have they impacted or changed anything in how you deal with product?
Oh, definitely. The major difference which I am seeing as a product owner — because I work very closely with the business, I try to understand their pain points and try to provide them solutions. There are a lot of situations where, when I am assigned a new product, from the discovery of the product to understanding the pain points there, previously it used to take like five, six calls with them. But these days it's not, because meetings are recorded — I can transcribe, the transcripts are there, the AI tools are there which can translate it for me.
Also, the big change which I'm seeing is, because I don't have to scrape down the whole company documentation to understand the business — I can just ask questions. It's a funny thing. It's like having AI as your co-worker with you, who has all the questions, all the answers, but it won't judge you for stupid questions, so I can ask all the stupid questions to it.
So rather than having to talk to all the stakeholders...
That's correct. Yes.
But you really only have to talk to them once, because everything is recorded, everything is written down, and then everything can be synthesized together. So it makes up for — what for you? Why would you have to talk to them eight times before, and only once now?
Exactly. I don't need eight times to talk to them to understand their business. We can directly jump to understanding the pain points and then solutioning it. So what is already existing is pretty easy to understand, and what we need to build in the future — those kinds of conversations we can jump to. It has fast-forwarded the whole process.
What kind of tools are you using for this?
I'm using Copilot — that's company-built. Personal life, I use ChatGPT, Perplexity.
Any Claude in there? Mostly Perplexity?
Yeah, that's it. Because my work is not on the purely coding side — Claude is very good with coding, and the technical perspective, the architecture, whereas Perplexity and ChatGPT are more generic.
What do you use ChatGPT for versus Perplexity?
There is this difference which I found: Perplexity also scrapes the whole web for you, so it will give you the sources and the source of truth. ChatGPT gives you its own perspective, so it will understand if you want to listen to the truth or not, or if you're in a bad mood or a good mood, and it will answer you accordingly. So between the two, I go to Perplexity more.
So you feel like ChatGPT is reading your mood and adjusting its answers accordingly?
Oh, yes. The way I ask questions — it's all query, right? The way it's answering me, and if I'm like, "You are not saying it, telling this right," or something, then it will start correcting: "Oh, you're right."
So you trust ChatGPT less than Perplexity?
Yes, because Perplexity scrapes down the whole web for you. It won't adjust its responses according to your mood.
Are you a little bit worried that the amount of content being generated for the web right now is more AI than human?
That is true. That is true.
You're excited about a bunch of stuff in some of these changes — what's keeping you up at night though?
I have answers to all the random questions in my head, and I can just ask. I can just text-prompt it and it will scrape the whole web for me, give me the opinions, give me the perspectives. Another example — I was watching a random horror movie last night, and I couldn't understand the ending. It was very iffy for me. I just texted, and I understood the whole opinion of the movie. So it's very convenient. We have a lot of knowledge in hand.
Some people say that AI is making people dumb, but I feel it's not true. It is making me feel smarter, because I know a lot of things now. It's in my hand. I can just search. So I have a different opinion on it.
Is anything scaring you?
Not right now, because with my position as a product owner — as I mentioned, I understand the pain points and talk to the business, really talk to the business more often. The product owner position is not going anywhere, because we need empathy, and solutioning the pain point is also from the human touch, because we also need to understand the architectural perspective and the limitations. So with respect to my role in the industry, I am not scared. But I need to keep up with the AI, keep up with the technology. I don't want to go obsolete. I'm really interested in learning the AI ethics side of it.
But what about everybody else's job? Oracle just laid off 30,000 people between the US, Canada and India. Facebook's laying off 15K. Are you worried about other people's jobs or anything?
Yes, I think that is scary, that is really scary. I have seen a lot of people who are not in software industry, but in semiconductor industry or any other industry, and they are able to write code. I can write code today — I'm from the business side, but I can write code today. So that's a little bit scary. People, companies, are laying off.
What are you most excited about right now?
Right now, with respect to technology, I want to see where this goes. But I also feel everybody needs to run with it, adapt with it, so that they don't go obsolete in the market.
Do you have any friends that are not keeping up with it?
No, I get to learn a lot from my friends. I have friends who are more in the product industry, so they are trying to keep up with it. I'm really scared of the students especially, because they are not getting any jobs. I have a lot of relatives who are not getting any job because of the AI disruption — companies are not hiring.
I guess that's the question — does your department hire, and do they hire junior people, or do they just not hire right now?
They do hire, but the hiring has become lesser. The job which three people would do, one person is able to do, because of the faster retrieval of data — we can retrieve the data as soon as possible, read the data. I am able to keep up with it because I just got a new role assigned, and I'm able to keep up with it just because I can query Copilot, understand the documentation, understand the business, and ask the right questions in the meeting. So I can prepare really well in the meetings. I think that's the difference.
Be part of the
conversation.
Whether you're a CTO who wants to be featured, a company looking to sponsor, or an engineering leader wanting a seat in the room — there's a place for you here.