is now

8 minute read · Published August 15, 2024

Claire Vo: The CPO who automates herself with AI

Latest Update September 10, 2024

“I couldn’t have done this before LLMs. It’s all literally thousands of times faster now.” That’s what Claire Vo told Every’s Dan Shipper about ChatPRD, the six-figure side project she builds next to running LaunchDarkly’s product and engineering orgs. 

If I taught a course about product management careers, Claire’s resume would be the example. She’s been a designer, engineer, acquired founder and three-time CPO. 

But she doesn’t rest on her laurels. Claire’s nights and weekends are spent on ChatPRD, which she describes as an on-demand Chief Product Officer. This might prompt a question:

Why would a Chief Product Officer automate her own job with AI?

It’s always great to learn from successful product leaders, but Claire Vo feels a bit like a time traveler. She sees the future of product clearer than most people I’ve met. That’s why I interviewed her about the future of product management and making a career in it. 

“All roles are gonna change, why not this one?” she told me. “People should be changing already. And I think if you haven't started down that journey, you're going to be too late.” she continued. 

Here’s what that means in practice - and how the early stages of PM automation look.

ChatPRD: The perfect startup?

Today, ChatPRD is a product manager copilot - it saves product people hours every day. Claire told me she often hears that ChatPRD customers now do in an hour what used to take a day. She views conventional PMs as rate limiters in organizations. ChatPRD intends to change that dynamic. 

ChatPRD landing page

But ChatPRD was never intended to be a company. Yet it could be a perfect case study in a book on building startups. It started with Claire solving her own problem. Writing PRDs is repetitive and takes a long time. So when ChatGPT came out, Claire found it passable for writing PRDs.

This kind of low-stakes experimentation often leads to discoveries that become successful companies. The same happened with Command AI: We initially built a cmd+k widget for our own product before pivoting into doing it as a service.

Back to Claire. As she shipped great PRDs quickly, her colleagues asked her what had changed. Once Claire told them her secret, they demanded her prompt. 

This is a great sign for early product-market fit: If you’ve solved your own problems and others want that solution, there’s a good chance it’d be a successful product. When OpenAI’s GPT store launched, it was the perfect launchpad for ChatPRD. 

What started as a ChatGPT prompt could now be a custom GPT with custom instructions, context and knowledge. The GPT spread within her company and beyond. 

This is a great sign of early PMF: If your product spreads without marketing, you have some measure of product-market fit. This spread was enabled by Claire following startup strategy perfectly: Initially solved one problem (Write PRDs) for one persona (product managers in mid-to-large companies). 

It’s always easy to reverse-engineer these things, but they’re often more random than they appear looking back. From what Claire shares in podcasts, she just wanted to share something cool she made with colleagues, with no secret master plan.

But an exited founder can’t shake the pull of PMF. Claire knew she could charge for this, but kept aspirations modest: Her goal was for ChatPRD to pay for a glass of wine a week. 

So ChatPRD got its own home at ChatPRD.ai, complete with a paid plan. Today, ChatPRD is far from paying for a glass of wine. It’s much closer to paying for a vineyard. 

ChatPRD kept growing: New features, a team plan, new team members — the classic startup playbook. Right now, the budding startup does six figures in ARR - and Claire’s ambitions have changed from paying for a glass of wine to paying for her kids’ college. 

You (and the numerous VCs who must be in her DMs) might wonder when Claire will quit at LaunchDarkly to build ChatPRD full-time. She’s not planning to. 

This is the power of AI-augmented builders. 

How AI augments product people 

Claire has many skills: She’s worked in marketing, been a designer, knows how to code and deeply understands product. 

That makes her the ideal solo builder. But even then, building a six-figure business next to a busy leadership job used to be basically impossible. To ship at the level Claire ships, you would’ve needed multiple engineers and a designer. 

AI isn’t only powering ChatPRD’s backend, but also Claire’s process. Even if AI doesn’t do most tasks start-to-finish as good as a human, it’s great at giving you a basis to work with. This accelerates most processes. 

There’s also another underrated aspect: It’s always there. ChatPRD may not be a Claire Vo-level CPO, but it’s always there and responds instantly. ChatGPT may not be a Hemingway-level writer, but it’s always there and responds instantly. 

Claire Vo isn’t the only builder who can now accomplish what used to take 5 people. And if something that used to take 5 people now takes 1, what does that mean for us product people? 

Claire also thinks the PM role will evolve. That’s because anyone now can access a “pretty good” level of many skills. Take illustration: AI might not generate illustrations like a great illustrator, but it lets anyone generate pretty good images. 

The same is true for writing, coding and much more - and it’ll only be true for more skills. When I asked Claire, she believes this means the PM role could develop into one of two directions.

The generalist PM

Product management has always been a generalist profession. You might do a dozen things in a day - and there are few clear skills every PM needs.

But in most companies, there was a clear boundary: PMs are managers. They don’t code or design. PMs orchestrate. 

AI could broaden the PM’s skillset: A PM can now prototype faster. That’s because they’ll have a pretty good engineer and pretty good designer (AI) who can instantly produce an output and act on feedback. 

This would make PMs more autonomous and elevate them to “product makers”, where engineering and design either start with the PM’s prototypes or jump in deeper. 

This is similar to what Claire did herself: ChatPRD started as an internal prototype which she only fleshed out later - and wouldn’t have been able to ship part-time without AI.

In the future, it’s easy to imagine an AI agent reading a PRD/Spec and coming up with a basic design or V1 of the code. It may still not be perfect, but the product designer might become more of a product maker, with more direct control over the builds. 

Tools like Command AI could also play into this: When it's easy to build in-product widgets and help your users, you need fewer people to ship a user experience that used to take whole teams weeks.

This would change the skills you need as a product person: Instead of aligning people and writing a lot of tickets, it would become about taste and vision. 

2 - Get rid of (most) PMs? 

If option one was PMs doing more engineering and design, option two is the reverse: Engineering and design doing product management. 

This would mean that design and engineering use AI tools to synthesize communication, assign tasks, give feedback etc. because the activities PMs typically do are largely automated by AI. 

It’s easy to see how tools like ChatPRD sketch this future: If PMs can already do in an hour what used to take a day, it’s easy to see how a design or tech lead might take over these aspects of product management.

Claire told me she believes this could mean there being fewer PMs, where not every product gets its own manager. This would raise the threshold for hiring a product manager for complex products. 

What skills do PMs need to learn to survive the AI age? 

Claire has the perfect background for a PM: She’s done both design and engineering. Not all product managers have a skillset that diverse. In the past, it felt like knowing design or engineering was nice, but optional. 

Will that continue to be true? I asked Claire. 

“Technical is a funny word. It can mean a couple of things. It can mean I understand computer science, it can mean I can code or it can mean I can build things. I think we need to disambiguate when we’re talking about being “technical”.

Code still has a mystical quality to non-technical people like me. But as Claire describes it, it seems obvious: There are levels to this game. 

Claire believes being a PM won’t require computer science degrees, but will require being able to build. This will be driven by two forces: 

  • AI enables more people to build things 
  • Companies will expect more product people to build things 

This will mean that the skill set of product managers will need to become somewhat more technical, even if it’s just in the “knows how to build things” way. 

So, what do we do about all of this? I asked Claire what she’d advise a young, bright person trying to break into product. 

How to break into product in the future 

Claire broke down her advice into two categories: 

  • What she’s always advised 
  • What she’d advise because of AI. 

Let’s start with the first, more general piece: 

“Find companies where you actually have something specific to add or an opinion. Reach out to those companies directly. The spray and pray approach to getting a junior/entry level job in product has never worked. And so the best thing that you can do is work on your network. And the best way you can build your network is add something of value to a hiring manager. And the best way you can add something of value to a hiring manager is get to know the company and the product and offer feedback or advice or ideas. So I think that's always stayed the same.” 

As Claire says, this has always been true, except for people with a perfect resume. But Claire believes AI means you should go a step further: 

“I would advise product managers to take that a step further and actually build something for the companies that they're interested in.”

That’s not just an abstract idea. She says this is a common ChatPRD use case: Job seekers create artifacts for companies they’re applying to. This could be a spec for a product improvement or a product strategy for moving the company forward. This helps people stand out in application processes. 

Right now, this is a bonus, something cool to stand out. But Claire believes this will become table stakes: 

“I think you'll be asked to provide not just content about yourself, but ideas about what you'll bring even ahead of having a formal interview. So that's one of the things I would advise folks to start thinking about. I think the bar is going to be high there.”

If you really want to stand out, Claire says, she has two words of advice: Build something.

She believes that you used to demonstrate thinking, ideas, frameworks, etc. But as that becomes table stakes, showing an actual prototype will become more important. 

“That would certainly cut through the noise with me. It would work on me if someone was trying to break into product”. 

This isn’t pure theory. As a CPO, she knows. The perfect example might be how she hired her first employee, the head of growth at ChatPRD, Alisa Haman. She got Claire’s attention by making a video about ChatPRD. 

As Claire told me: “She actually created an artifact that was extremely useful. It showed that she had thought about the product, that she had opinions. It also showed off skills that wouldn't have shown through on a resume. For example, video editing skills, marketing skills, growth skills. 

It made it a no brainer when I was hiring to say this person knows my product, they care, they have ideas for how to make it better and I can see their hard skills because they have demonstrated them ahead of me.”

Copy icon
X logo
LinkedIn logo

Zero fluff.

Our stories 🤝 your inbox

Email Icon

Up Next

Mastering adoption metrics: No boredom included

Never miss another
fabulous article.

Weekly digest in your inbox.

Email Icon
Continue Reading

Up Next

Command Palette: Past, present, and future

Read Now

What's hot