AI-Powered Patent Claim Drafting. My New Workflow.

00:00:00: You have just uploaded an entire invention description, boom, one click of a button and you now have like a thirty page description on your desk.

00:00:07: It looks super convincing and you are thinking, hmm, is it good to go?

00:00:12: And that's a huge problem because you didn't actually write it yourself.

00:00:15: And even more importantly, because you didn't actually think it.

00:00:19: Instead, you've become what recent research calls an intellectual tourist in your own work.

00:00:24: You are visiting ideas, but you are not inhabiting them.

00:00:27: And by letting a one click robot do work you have unknowingly stepped into the role of a middle manager of your own thoughts.

00:00:34: You're now just validating a robot's statistical predictions.

00:00:38: So let's unpack why this is such a huge problem.

00:00:40: Researchers call this outsourced reason.

00:00:43: We are entering an era where we are risking delegating our fundamental reasoning to algorithms, which then alienates us from our own craft.

00:00:52: And there are three psychological risks why we stop thinking critically ourselves.

00:00:56: Risk number one is the trap of satisfying.

00:00:59: Satisfying is a strange term, a combination of satisfying and suffice.

00:01:05: The human psychological tendency to accept NAI solution and AI created a text that merely meets, let's say, a minimum threshold of apparent correctness rather than the human searching for the optimal strategic path.

00:01:21: The second risk is the filled page problem.

00:01:23: So back in the days when we were writing our texts ourselves, we had the writer's block sometimes in the sense that you had a blank page in front of you.

00:01:35: But now the problem has shifted to staring at a page, AI has filled for you and you are wondering whether you can actually agree with it.

00:01:42: And risk number three is mechanized convergence.

00:01:45: And this is the funny property that AI somehow gravitates toward the statistical average, producing a smaller, more homogeneous range of ideas, which then eventually results in average patterns.

00:02:00: Let me share you my workflow, how I'm actually solving this problem.

00:02:05: So I'm in the middle of a conversation with my AI assistant here already and I want to draft the claims for a new patent on a smart dosing system, short context.

00:02:16: So imagine you are at the supermarket and you want to buy some rice or some other bulk item, right?

00:02:23: You can go to a self used dispenser, right?

00:02:28: You can put your container under it and then dispense the rice into your container.

00:02:33: Now, because, you know, high precision trade certified scales are pretty much expensive to install at every single nozzle.

00:02:42: the dispenser has only a very cheap non-calibrated scale which is just fine to get the amount roughly right.

00:02:49: Now then later when you go to the checkout and when you are actually paying there the supermarket has a high-end certified master scale so to say and this one then calculates the final price down to the gram.

00:03:03: So that's the invention here.

00:03:05: and going back to my claim drafter So I've already uploaded my invention description and now instead of jumping straight to the drafting of patent claims I've programmed my assistance in such a way that first of all, conducts a critical conversation about the invention in order for us to reach a common understanding in the sense that this is not a user pleasing servant, but it's actually a tool for thought, which helps me think better.

00:03:37: So after I've uploaded the documents, it starts with the first clarifying question.

00:03:44: It says I see the invention covers a broad system with multiple components.

00:03:48: We have dosing units.

00:03:49: We have the reusable containers.

00:03:52: We have this central weighing station and some other things in the invention description.

00:03:57: And now the assistant is asking me the question, what's actually your strategic protection goal, right?

00:04:05: Do you want to protect the entire ecosystem.

00:04:08: Do you want to focus on the in-store dispensing and weighing process?

00:04:13: Do you want to focus on the container circulation logic?

00:04:16: So in this case, I would like to focus on the software that controls the process and connects the two weighing tasks.

00:04:36: Let's let's call it that way.

00:04:39: And you see now I'm, I'm entering a conversation with my AI.

00:04:43: And now it's asking me the next clarifying question.

00:04:46: What do you consider the technical scope of the software?

00:04:49: Let's do the full software chain option D. And now the conversation goes on.

00:04:55: Who is the intended infringing actor?

00:04:57: Hmm, okay, I would say in this case, it's actually, I think my client in this example would not want to sue the retailer because that's their customer but they want to.

00:05:12: So the software vendor, option C, who develops and licensed similar orchestration software.

00:05:19: So let's go for option C here.

00:05:20: Okay.

00:05:21: And what are we talking about now?

00:05:24: What do you consider the core unique principle that distinguishes this software approach from existing bulk dispensing systems?

00:05:32: Okay.

00:05:33: So now that we have, you know, the system boundaries and the infringing actor, now we are diving into the real technical US And for this example, I would go for the two-stage weighing split.

00:05:46: Of course, if the answer options provided by my assistant here are not on point, I can always go into a deeper discussion here.

00:05:55: I'm just skipping that for the sake of this example now, right?

00:05:58: Now we are talking about technical advantages.

00:06:01: Here, my assistant is suggesting some of them and I can now, you know, drill deeper into that.

00:06:08: But for now, let's just say, let's go with these.

00:06:15: Now, yeah, we have the technical advantages.

00:06:19: Here's as a next step.

00:06:22: Now, the question what's the actual hierarchy of the protection, so which are, you know, the fallback.

00:06:30: positions.

00:06:30: Now my assistant suggests a list of additional features from the invention description and asks me to rank them.

00:06:39: So in this case, I would just say BDC.

00:06:43: for the sake of example, and there we have it.

00:06:45: So now the hierarchy is set.

00:06:48: Now my assistant is kind of poking into the technical and integrated details, right?

00:06:54: And yeah, we can have a discussion about anything that's unclear or missing.

00:06:59: In that case, I would just say make up your own.

00:07:04: mind to not overstretch the example here.

00:07:08: And now the AI assistant and I have collected enough information and structured so that we can actually come up with a consistent invention summary.

00:07:20: This is structured and this is basically the first step in my workflow.

00:07:25: understanding the invention really thoroughly together with the AI.

00:07:29: And I would now continue to go through the whole process, you know, identifying the prior art, then figuring out the distinguishing features and their technical effects, defining the objective technical problem.

00:07:43: And then we would draft claims together, my AI assistant and I. And I hope this shows you what I mean when I say I'm using AI as a tool for thought.

00:07:53: So to make it crystal clear, the one click robot kind of stands between the attorney and the work, right?

00:08:00: While a tool for thought actually sits next to you and elevates your own reasoning.

00:08:07: Now, as we've seen in the demo, my AI assistant is based on three design principles.

00:08:13: Principle number one is material engagement.

00:08:15: The assistant forces me to, you know, touch the clay.

00:08:19: The AI is asking me uncomfortable questions that force me really to personally dig into the technical details of the invention and into the strategic questions.

00:08:31: Who is the infringer?

00:08:32: What's what are the system boundaries?

00:08:33: What's the protection hierarchy and so on rather than, you know Just dumping the inventors notes into the AI and hoping that there will be a good draft coming out of that.

00:08:45: principle number two is productive resistance.

00:08:47: So the AI assistant is acting as a provocateur to provide productive resistance, you know challenging my assumptions and asking really critical questions to show my own judgment in the end.

00:09:02: Principle number three is scaffolding metacognition.

00:09:05: This means that the machine, the AI assistant handles the manual data space, the text production in the end.

00:09:12: See the invention summary we were drafting.

00:09:14: And if you use the claim drafting assistant for the whole process, of course, the final claim set in the end, which will be drafted by the AI.

00:09:21: So the machine takes over the text production so that I have the mental headspace for more critical thinking and for the bigger picture strategic decisions.

00:09:31: In the end, it's all about the human factor.

00:09:32: It's really about what I would call reclaiming the ghost in the patent in a highly digitized, streamlined world as we have it today, right?

00:09:43: The goal is really to inject the human intention back into the work to avoid producing AI slop.

00:09:48: So the ultimate goal of my paradigm of using AI is in the end, better answers.

00:09:54: I think that higher quality work is the result of asking better questions, not just seeking efficiency.

00:10:01: In fact, it's the other way around.

00:10:03: So the goal of my AI patent drafting assistant is better thinking and efficiency is then a nice byproduct.

00:10:13: So how can you put all this into practice?

00:10:15: Well, I've put some helpful resources into the description.

00:10:18: Check the links.

00:10:19: There are free prompts.

00:10:20: There's a link to my claim drafting assistant, which I just showed you and also to my seminar.

00:10:26: And if you want to fully understand the tool for thought paradigm I was talking about, I have another video on that that goes deep into that topic, which I'm also linking.

00:10:36: So go ahead and watch that next.