Paragon’s founder and CEO on building the future of patent drafting and AI for IP at Tradespace — an interview

Last month, Tradespace announced its acquisition of Paragon Patents, the AI-powered patent drafting tool founded by AR Bhatti and his team at Princeton. The acquisition didn’t just add new capabilities and features, it brought together two companies that think similarly about invention, IP strategy, and how AI should actually work with real people. I sat down with AR to show our community what I mean. 

Alec: AR, how does our story begin?

AR: So Paragon actually started as a technical experiment with Princeton’s Office of Innovation. GPT-3 had just come out, and it made me recall how incredibly tedious the patent drafting process was. I started thinking: would AI have made my life easier back then? Then I heard from one of the leaders of our Princeton entrepreneurship community that drafting patents was also a source of pain at the Office of Innovation.

Claire, Ethan, and I had been looking for something to work on together anyway, so we started working with the Office — their disclosures, correspondences, patents — and talking to lawyers and other TTOs to figure out what we could build. At some point, the Office told us that we had to meet Tradespace — that it was a rare legaltech product they actually enjoyed using and had completely changed how they ran their IP and tech transfer process. We’d actually just come back from AUTM and were starting to get real traction, so the timing felt right and we reached out.

Alec: You demoed the whole platform on our first call, and I remember being struck by how passionate you are about patent drafting. Not in a founder pitch way — though you’re clearly good at that too — but it was obvious you care deeply about the problem itself.

AR: I really appreciate that. What I noticed on that call was that you asked genuinely smart questions about how we were thinking about the problem, not just our tech. I’d already been impressed by how end-to-end Tradespace is, but by the time we wrapped up I knew this wasn’t going to be one of those conversations where everyone nods politely and moves on. Though honestly, at that point we were still just thinking we were going to do a partnership.

Alec: From my M&A advisory days, I’m always looking for those 1+1=3 plays. You had the enthusiasm, you were willing to iterate on feedback, and you showed real tenacity in following up. But beyond that, what you’d built was genuinely valuable.

Your approach to patent drafting — keeping humans in the loop, doing generation step by step, full traceability — that’s what makes attorneys actually accept AI in this process. How did you land on that?

AR: I think it came from Paragon not being my first time building. Ghost Pacer, my first startup, made AR glasses that gave runners a holographic racing partner. We had to iterate on that product a lot, and hardware iterations are brutal. So with Paragon, we were obsessed with getting customer feedback at every step and quickly evolving the program. And talking to over 200 lawyers taught us something crucial: not every part of patent drafting is ready for automation.

I always use the analogy that AI is like those motorized walkways at airports. They let you go faster, but that’s only helpful if they’re taking you to the right gate. And you wouldn’t put one at security, right? Same with patents. So we focused on keeping the lawyer in charge, not just showing them a finished product and saying, “Here’s what we did.”

Alec: That aligns with something we’ve learned too — attorneys don’t respond to AI for AI’s sake. It has to fit into existing workflows and actually make their lives easier.

AR: Exactly. One of my mentors talks about this distinction: there are products that depend entirely on AI, and then there are products that existed before but are made more powerful by it. We always thought of Paragon as patent drafting that’s AI-powered, not AI patent drafting. The point isn’t the AI. The point is making someone’s life better or easier.

About half our features — some people’s favorites — don’t use AI at all. Things like updating figure references. We built it because people kept saying: I hate doing this manually. No AI was needed to solve it, but the feature still mattered to us.

Tradespace is similar. You’ve been around longer than the LLM boom, and you made an intentional pivot to incorporate AI into key parts of your process. You didn’t just do it because that’s what you’re supposed to do now. That’s the difference between building something real and chasing trends.

Alec: Let’s go back to Ghost Pacer. IP was central to your role there and the company’s success, right?

AR: Oh man. I filed ten patent applications over Ghost Pacer’s life and drafted about 95% of it myself. It was brutal work, but necessary — saved us at least six figures because I had the background to do it. And it paid off. The driving force behind Peloton’s acquisition was our IP portfolio.

Alec: That’s incredible. What’s your advice to other founders on IP?

AR: Spend an hour mapping out what’s really novel about your creation — both today and five or ten years from now. Where does it fit in that larger landscape? If it’s significant and needs to be protected, start exploring your options.

And know it’s not just about filing patents. It’s about understanding what makes IP valuable in context. Understanding where the industry’s headed, who else might be interested in what you’ve built. That strategic thinking — building IP that’s future-proof, not just what covers you today — that’s what caught Peloton’s attention. That’s what made the difference.

Alec: That echoes what another exited founder, Jerry Ting, has said about IP changing his startup’s trajectory. But you started patenting long before Ghost Pacer. Take us back.

AR: My first invention was a long-range wireless charger. You’d plug a little box into an outlet, and it would beam power to your phone across the room. I filed for that patent when I was thirteen, and the process was long and expensive. So Paragon wasn’t just born from wanting to help Princeton. It also came from living through how painful the patent drafting process actually is.

Alec: Amazing. That firsthand experience — still in college but with this deep patenting resume — that’s a big part of why we’re so excited to have you on the team. What are you most looking forward to about joining Tradespace?

AR: I think we’re going to see major changes in IP over the next five years. The USPTO is going to have to rethink patent review. We need to figure out how prior art search works when AI is involved. These are big structural questions. The companies that figure out how to work with AI thoughtfully — not fighting it or ignoring it, but actually integrating it into how they operate — those are the ones with real advantages.

I think Tradespace’s culture is what’s going to let you soar on this front. When we visited San Francisco, it was immediately clear you’re already thinking this way. People actually want to be here. We felt heard and listened to. I saw how you iterate, how you make decisions, the commitment to excellence. And having a CEO who actually listens to their engineers? That’s rare. That’s what gives you the advantage. Customers feel it. You’re not just doing this — there’s a real mission. That’s what lets you integrate AI thoughtfully, incorporate feedback fast, stay ahead of what’s coming.

Alec: When I watched you, Ethan, and Claire’s reaction to our culture and how our people operate, I realized we weren’t just talking about a product fit — we were looking at people who wanted to build something together long-term.

AR: We’re genuinely excited to be part of that mission at Paragon — I mean at Tradespace! Sorry, I need to stop saying Paragon!

Alec: Good thing you’ll have plenty of time to get used to it.

AR has joined Tradespace as VP of AI Products & Patent Drafting. In the near term, he’s focused on integrating Paragon’s capabilities into the broader Tradespace platform. After that, he’ll be helping us evolve with the principles that made Paragon so great: keeping inventors and attorneys in control, building only what they want and need, and understanding that AI is a tool, not the destination. If you’re curious about what this looks like for patent drafting, join us for a group demo here or sign up for a 2-application free trial here.