Thoughts from the team.
Counsel can edit an AI drafted office action response in ArcPrime by marking a passage and saying what to change, or by typing in the draft. When a claim is amended, the arguments, the mirroring method, system and computer readable medium claims, and the dependents are updated to match.
Get the latest insights on IP strategy and patent portfolio management, straight to your inbox.
See how ArcPrime connects your portfolio, workflows, and business context to help your team make better decisions across the patent lifecycle.
An invention disclosure form in ArcPrime can carry questions only counsel see, marked for the company's IP team only or for outside counsel as well. Inventors never see those questions or their answers, and AI drafts a first answer from the submission.
ArcPrime's AI patent assistant answers questions about the record counsel has open, including the part of it in view, any selection and any filters. Questions can also draw on data from elsewhere in ArcPrime, such as finding claims across the portfolio that are similar to those in a continuation recommendation.
ArcPrime brings EPO, CNIPA, JPO and KIPO office documents onto each European, Chinese, Japanese or Korean case as the case is added or updated, drawing them from the USPTO's Global Dossier. On European cases, a focused view lists the office actions, replies and amendments.
Any AI claim chart and prior art search tool should be judged on at least 3 dimensions: 1. Did it find the evidence? 2. Is the reasoning sound? 3. Can you trust every citation?
A precedential USPTO decision and new examiner guidance make Section 101 rejections of AI inventions more winnable. What changed, and how to file.
Trademark management software for Madrid Protocol filings: ArcPrime holds designations and national filings in one family and shows coverage by class.
ArcPrime's patent docketing software tracks USPTO extensions of time on office actions, showing extensions used beside the final statutory date for each task.
ArcPrime's claim chart generator starts from a pasted list of competitor patent numbers, flags any not found, and keeps the claim charts private to the company.
How to measure outside counsel performance on patent work: 22 metrics across prosecution, claims, drafting and timeliness, weighted and traced to matters.
An invention disclosure workflow in ArcPrime saves each change as a labeled version counsel can compare or restore, with figures and tables kept in the text.
Patent budget forecasting in ArcPrime shows how a filing or annuity policy change moves next year's patent spend by month, against the baseline and budget.
Yes. ChatGPT can analyze a patent portfolio, but the depth depends on how the data reaches it: pasted claim sets get a solid first pass, while a live ArcPrime connection analyzes current families, pending deadlines, and disclosures still in review.
Yes. Claude, ChatGPT, and Cursor can all query your live patent portfolio today through the ArcPrime MCP server, so answers come from your current docket instead of a pasted export that goes stale.
As AI spreads across a large company, employees who were never hired to invent now produce patentable work in everyday tasks. Traditional invention disclosure processes miss it because those employees don't recognize their work as an invention. Capturing it means reading the trail of work they already produce, with a tool like an innovation listener, instead of waiting for a disclosure form.
Yes, putting patent data into ChatGPT can be safe, but only when the vendor clears the same diligence you apply to any custodian of confidential material: an independent audit, contract terms, and a written promise never to train on your data.
ArcPrime now builds a patent product taxonomy for every account: business segments and product lines with cited revenue on each node, so filing recommendations name the product line an invention protects and the revenue behind it.
Every ArcPrime claim chart now carries a strength score from 0 to 100 next to its High, Medium or Low badge. The claim chart strength score moves with the logic behind each mapping and the quality of the evidence, so charts inside a band can be ranked before a negotiation.
ArcPrime's AI summaries now name each patent's point of novelty, anchored to exact claim language, and carry it across claim charts, family comparisons, continuation recommendations and office action responses. Reviewers can edit any element and their version stays.
ArcPrime now runs claim level portfolio analysis on patents you don't own. Score an acquisition target's IP or run a light freedom to operate check.
ArcPrime's Innovation Harvester reads Jira, GitHub, Slack, and Google Docs for inventions that never became disclosures, and drafts the disclosure itself.
Upload one product document and ArcPrime screens your entire patent portfolio for claims that read on it, pairing your evidence with its own web research.
When handling thousands of cases per yr with a small inhouse team, my primary strategy for ensuring patent quality at scale was to pick trusted outside counsel. I had to trust they focused on issues that I cared about, and I'd spot-check the work. (I don’t think this changes with AI. Spot-checking just becomes checking every draft.) In any event, I always wondered: what problems were still making it into patents when nobody was looking? [numbers below...]
ArcPrime continuation recommendations now map each candidate to your products or a competitor's, so you rank filings by coverage, licensing, or enforcement value.
ArcPrime scores your patents by infringement likelihood against a target product, each with a percentage and a rationale you can defend.
ArcPrime turns a plain language question into live patent reports and dashboards, with the portfolio mapped to revenue for board and GC reporting.
ArcPrime pairs your outside counsel with AI office action drafting, so each response takes under two hours to review and costs up to 50% less.
ArcPrime's MCP server connects Claude, GPT, or Gemini to your patent portfolio with read and write access, so your AI tools act on real IP data.
There’s a lot of debate around whether AI can one-shot patent tasks or whether human-in-the-loop will always be required. Based on what our team has achieved, I believe it will be closer to one-shot than not.
A legacy IPMS migration teardown: the failure modes, why switching IP systems goes wrong, and how to move your patent data safely and be live in weeks.
Best IP management software, a spreadsheet, or a legacy IPMS for your patents? An honest buyer's guide to choosing what fits and protects your portfolio.
What in-house patent teams told us at IPBC Global 2026: build-vs-buy pressure, AI-native tooling, and outside counsel disrupting their own pricing.
Maintenance fees add up fast. Here are 7 patent pruning strategies that cut renewal costs without giving up coverage that matters.
Not all IP management platforms are built for how IP teams work now. Here are 8 must-have features in a modern IP management platform.
Choosing an IPMS is a multi-year commitment. Here are 7 questions to ask before buying an IP management system so you pick the right one.
The license fee is only the start. Here are 6 hidden costs of legacy IP management software that quietly inflate your total cost of ownership.
Tracking your portfolio is not the same as acting on it. Here are 7 signs your IP team has outgrown its tools and needs a modern IP management system.
No two IP organizations work the same way. Different industries, different team structures, different approval chains. Until now, most IP software asked you to adapt your process to fit the tool. We think it should be the other way around.
When an invention or application has been sitting in a stage longer than expected, ArcPrime automatically notifies the relevant outside counsel. No manual monitoring, no calendar reminders, no nudges from you required.
Your portfolio generates a lot of data: applications filed, prosecution history, maintenance decisions, competitor activity. Portfolio Memory connects those signals into guidance that arrives in your inbox every month, without you having to ask. As your portfolio evolves, the recommendations evolve with it.
We're rolling out Auto Prior Art Search so relevant references surface automatically, without anyone having to kick off the process manually.
Filing Score gives your team a shared framework for making filing decisions. One that's defensible, repeatable, and tied to real outcomes. Invention review meetings get faster. Stakeholder conversations get cleaner. The portfolio that results gets stronger.
ArcPrime is a sponsor of IPBC Global 2026 in San Diego, June 15–17. Meet us there to see what AI-native patent portfolio management looks like in practice.
I like numbers more than most people (as the "analytics" guy at Meta), yet I was always wary of patent quality scores, i.e., how good is this patent from 0-100? There's always a heated debate when discussing whether to introduce these to the team.
We're so grateful you're willing to share your experience. Reviews from customers like you help other IP teams find ArcPrime and make confident decisions. Your voice carries real weight, and we don't take that lightly. The whole process takes about 5-10 minutes.
15 years in patent work and there's one question that never leaves the back of my mind: how do I know which patents actually matter? In 2 weeks, I'm hosting a 1-hour California CLE approved webinar on how to get the most out of AI today to do strategic patent work.
One of my favorite parts of being in-house counsel was meeting with inventors. It was how I knew what the company was actually building and where the world was going.
After a few years in-house, I started to worry I'd lose my sharpness as a drafter.
Here are 10 prompts that’ll replace your patent attorney!! Jk, but these prompts really do help. They draft claims from disclosures, review application drafts, search for prior art, build claim charts, file continuations, and make portfolio decisions across thousands of patents.
A group of us IP attorneys are starting a community for inhouse counsel who want to learn how to use AI well in their daily work. AI is already transforming how legal work gets done, and execution is no longer the bottleneck. Improved results will not only show up for the individual via legal work product, but whole organizational structures & expectations may need to be rebuilt around workflows enabled by AI. For example, counsel may now need to spend time coordinating agents (adjust prompts, providing context) to do work they couldn’t do before.
We may be entering a world with dramatically stronger patents. The rate of innovation and filings may stay the same or accelerate, but the bar for non-obviousness could rise as models get stronger.
We analyzed ~9,700 asserted patents and found that enforcement is typically a mid life event, with a median assertion age of about 13 years. The data suggests that early prosecution choices and continuation strategy, not claim volume, are what most often determine which patents mature into real leverage.
Patent forward citations are often treated as a shortcut for patent value — more citations supposedly mean higher enforcement potential. But recent analysis of 4,300 US patents shows many asserted patents have surprisingly few citations, while highly cited patents aren’t always enforcement-ready. Could your portfolio be overlooking hidden enforcement leverage?
Patent family size and continuation strategy play a critical role in which patents are ultimately asserted. Drawing on 6,500+ US patent litigations, this piece explores what the families behind asserted patents reveal — and how early prosecution decisions can shape enforcement readiness long before litigation begins.
Analysis shows that patents ending up in court are far more likely to be acquired than invented, challenging the assumption that organic filing alone delivers enforcement-ready coverage, based on 6,500+ US patent litigations in the last 5 years. If assertion is so often bought on the secondary market — by both NPEs and operating companies — what does that mean for how portfolios should really be built?
What can the claims of 9,600 asserted patents from the last five years tell us about building stronger portfolios? This post uncovers surprising claim drafting patterns across industries and kicks off a short series using real litigation outcomes, not intuition, to shape portfolios from the patent up.
I analyzed 9,700 asserted patents from 6,500 litigations over the last five years and compared them to a control set of 10,000 randomly selected granted patents from the same period. Here's the insights I found most interesting:
As I reflect on 2025, I think about the 2 characters that changed my entire mental model of the world: O3. Released only 8 months ago, O3 is OpenAI’s first reasoning model that can produce genuinely useful legal work. ArcPrime's features, from claim charts to continuation recommendations, only work because of reasoning models. They are what give folks “tingles” on the future of the profession.
The worst thing about patents is the SLOW feedback loop. The field doesn’t evolve quickly, bad policies and inefficiencies linger, and there’s almost no room for real experimentation. You can file something today and not learn whether it mattered until long after everyone involved has changed jobs.
Building a patent portfolio is more political than people think. The obvious answer, "patent what's valuable", is not so straightforward.
When I first got to Facebook, I was dying to know how the world’s top companies actually used patents. Up to that point, all I knew was how to write and obtain high-quality ones — the craft, not the strategy.
I'm ecstatic to share that ArcPrime has raised $5.1M 🎉 in combined seed and preseed funding, led by early Datadog-backer RTP Global and Rackhouse Venture Capital, with participation from Leonis Capital, Atria Ventures, and leaders from NVIDIA, Meta Superintelligence, Apple, and Braze.
This year, most of our team met in Chongqing, China — the multi-level cyberpunk city. In Chongqing, uptown and downtown describe altitude, not direction. We surrounded ourselves with neon lights one day and raging waterfalls in Wulong Karst the next.
People tend to overcomplicate patent strategy. But it’s really simple: nobody pushes around the person with the biggest gun.
Most advice is a Rorschach test - it means different things to different people. Take the classic: “Do one thing well.”
Great to see everyone at IPO! It was a great experience meeting new folks from all over the industry. We did 2 things:
I've seen thousands of continuation recommendations. Our AI-generated recommendation beats all of them.
Most companies don’t know which of their patents are the best. And how could they? To find out, you’d normally need outside counsel to build claim charts costing 5 or 6 figures.
Super proud of the team for what has been a monstrous July! I'm most excited about patent charts and product charts. I've long wondered how to identify the most valuable patents in a portfolio at scale. Traditionally, folks used analytics such as number of forward citations, among a few other factors. While this is useful, and we are shipping this in August, the better proxy of value is if you can find live products in the marketplace using your patents.
Most ideas aren’t obviously great. But they’re also not obviously bad. They exist in this uncomfortable middle ground. This is where companies make their worst patent filing decisions.
A partner at a law firm once looked at something I wrote and scolded, “Do you know how to write a paragraph?” I thought I was better than average given I passed the bar and did well in some college writing courses. Apparently not. I had to relearn the basics: use topic sentences and break ideas into digestible pieces.
Many attorneys are hoarders. They'd pay any fee to keep a case alive until the bitter end. Ask them why and you'll hear the same thing: "Just in case!" or "But I've spent so much money on it!" Are these strategies or habits?
We’ve grown faster than expected since launching almost two months ago! There’s real hunger for 1) AI-powered workflows that make teams more efficient and 2) a modern take on IP software.
Slightly hot take: It's a great time to be a specialist. I used to think generalists had the edge. David Epstein's Range convinced me that breadth beats depth. I even switched from patent counsel to product counsel, hoping that exposure to different legal domains as a 'mini-GC' would make me a better attorney.
If you're looking for an easy 10% performance boost, minimize the friction of getting your thoughts to an LLM. Install the desktop app, learn the shortcut, or keep the tab open - whatever gets you feedback faster.
When choosing outside counsel, pay attention to the associates. As a former associate, I learned that they don't just support the process; they critically shape it.
The first time I saw true craftsmanship was at Fish & Richardson. Seeing Hans Troesch's work was like seeing a new primary color. Hans is a once-in-a-generation craftsman who redefined "excellence" for me.
Hello, world. I've been lurking for 20 years. But the problems that excite me now won’t let me stay there.