I spend my days connecting with startups and founders who are exploring partners to help them create the first versions of their product. At this early stage, that means helping them find product–market fit, validate ideas, and make the right calls on technology, design, and go-to-market strategy.
Our approach has always been grounded in a user-first mindset, even as AI tools like ChatGPT change how we work. Lately, more and more founders have been sending us project briefs written entirely by AI. These can be a great way to share your vision quickly, but they are not always the best source of truth or an appropriate first phase when you are about to make a big investment in building your product.
Here is where ChatGPT is helpful, where it is not, and how to bridge the gap.
The upside: benefits of using ChatGPT for a project brief
ChatGPT’s biggest strength is speed. In minutes, you can have a draft project brief that might have taken hours to pull together. For busy founders, that can be a lifesaver. It is also a great tool for brainstorming. If you are stuck, it can spark ideas, suggest features, or help you frame your problem in a fresh way. This is very helpful if it’s your first time exploring or building software, or you have a non-technical background in specific elements.
It is also good at making things clear and concise. A brief that is well-structured and easy to read helps potential stakeholders, investors, and outside groups, like agencies or developers get on the same page with you quickly. Even if you already have a draft or elements of one, ChatGPT can act like a second set of eyes, pointing out gaps or areas that need refining and then proposing some ideas.
By handling some of the early legwork, AI can absolutely save you time and money upfront, and can be a useful tool for preparing you to begin conversation with development partners and your target user base faster.
The downside: where ChatGPT falls short
The catch is that AI cannot bring human context or lived experience to the conversation. It does not deeply know your market, your users, or what will make your product stand out. Without depth, briefs can come out sounding generic or miss important details.
If you are in a niche or fast-changing industry, AI’s suggestions may feel outdated or off-base. It also leans heavily on existing patterns, so you are less likely to get truly original, disruptive ideas that help you win in a competitive market. And while a neat, well-organized brief looks great on paper, it is not the same as one built on real-world insight from launching and iterating MVPs and will likely raise a few flags with investors. The issue isn’t using ChatGPT to generate ideas - the concern is having only that as an input, and limiting real-world research as a supporting medium.
For example, if we entered a brief summary of TrueVisa into ChatGPT and asked for a project brief that was ready to be shared with potential partners for building an MVP, here is what it broke down for Target Users.
- Immigrants and families seeking affordable, accurate support with USCIS forms
- Community organizations looking to refer clients to trusted, accessible tools
- Attorneys interested in scalable ways to serve more clients
While this is a helpful starting point, here is how thoughtbot would take this deeper.
Where the human advantage comes in
This is where an experienced product team makes all the difference. For example, our team can take what you have, whether it is an AI draft, a bullet list, or just an idea, and turn it into a validated, strategic blueprint for success using a sound approach. Our process is rooted in user research, market insight, and years of building and launching products.
We help you identify your target audience, their critical path, prioritize the features that matter most, and proceed with implementation confidently. Along the way, we share the reasoning behind every exercise so you are equipped to grow and evolve your product after launch.
Going back to our TrueVisa example, we started that MVP effort with Discovery work, which included conducting 1-1 user interviews. We ended up interviewing participants from 11 different countries, who applied for 5 different types of visas, and had a good mix of those who used attorneys and those who didn’t. We also took time to speak to attorneys about their experience supporting individuals. These interviews helped us learn:
- Users would recommend this product -> highlights strong opportunity to organically grow reach
- Users do not need site translations -> allowed us to trim a substantial piece of effort from MVP
- Users feel more confident with attorney assistance -> we confirmed they found value in an automated solution, but needed messaging that ensured submissions were backed up by attorneys OR a way to engage one
A useful tool but not the only tool
Bringing to a close our TrueVisa ChatGPT vs. thoughtbot example, ChatGPT did a good job producing a concise vision. That vision was more thoroughly validated through research and by doing so we not only understood more about target users, but it also impacted the strategy of what we were building for the better.
If ChatGPT got you to the starting line, great. Now it is time to make sure your actual roadmap is ready to lead a real build that works for your users and your business.
If you have an idea for a solution or a WIP Project Brief, let’s talk about turning your vision into a market-ready product.