Time versus money in foreign IP registration: how AI helps save both

For most of my 13 years as a patent attorney, I’ve seen the same story unfold again and again. Whether it’s a two-person startup or a multinational corporation, the moment a company starts filing patents or trademarks abroad, it faces a familiar dilemma: pay more to have someone manage it for you or save money by managing dozens of local firms yourself. There’s no easy choice – only two kinds of pain.

The first path looks comfortable at first glance. You hire a ‘prime’ contractor – a large international firm that promises to take care of everything. For example, a US company hires a US law firm, which then subcontracts filings to local patent attorneys in ten or fifty countries.

You get peace of mind: one invoice, one point of contact, someone who claims to handle the chaos for you. But hidden inside that comfort is a huge premium. You’re forced to work with whoever your intermediary selects, often without knowing how good they are or how many layers of subcontractors exist. Decisions get made without your knowledge. Timelines stretch as messages bounce from you to the intermediary to the local counsel and back.

And you pay dearly for that privilege. I’ve seen invoices where an Indian patent attorney’s full office action response costs $800 – and the ‘management’ of that response by a US law firm costs $2,000. No substantive work; just forwarding emails. Filing a trademark in China costs $150 in total, but by the time it passes through a Western intermediary, it becomes $1,500. The maths is absurd, but it persists because the system rewards inefficiency.

Realising the above, many clients try to go direct. They start contacting local firms in India, China, Japan, Australia – everywhere they plan to file. It feels empowering at first. You’re saving money, getting direct answers. But very soon, you drown.

Fifty emails for fifty countries. Fifty different ways of quoting, invoicing, and communicating. One attorney uses spreadsheets, another sends PDFs, a third asks for scanned signatures. You now need a system to track it all – or, more likely, you need a paralegal whose full-time job becomes managing this chaos. Even a junior legal assistant can cost several thousand dollars a month.

So you’re back to paying too much – either for other people’s hours or for your own. And regardless of the model, communication suffers. In the first case, it’s filtered through layers; in the second, it’s unstructured and scattered. Both paths waste time, money, and energy.

For decades, the intellectual property (IP) industry has tried to digitalise this mess. “Software-as-a-service”, they called it, but under the hood it was still service. Platforms that helped submit to multiple countries, track deadlines, or exchange messages with local counsel appeared, but they didn’t remove the human bottleneck.

You still had to talk to local attorneys manually. You still paid human hours to coordinate, check, and follow up. The tools reduced startup friction but didn’t fix the core inefficiency. The cost of management remained, just repackaged.

When AI became the next big thing, many hoped it would finally untangle this web. But most ‘AI in IP’ efforts focused on the glamorous side: drafting patents, writing office action responses, and conducting prior art searches. That’s high-skill work – but it’s not where most of the time is lost.

The real time sink is operational: reading and sending thousands of emails, updating deadlines, and moving data between systems. These are repetitive, well-defined tasks that cry out for automation, but they involve unstructured communication with external partners who all work differently.

I’m working with a large firm managing 20,000 cases. Two operations managers handle 18,000 domestic cases, but four are needed for just 2,000 foreign cases. Why? Because cross-border work involves external providers – each with its own rules, formats, and habits. It’s not that humans can’t do it; it’s that they have to do it manually.

This explains why the market has barely touched this part of the problem: it’s messy, fragmented, and full of exceptions – everything machines hate.

So what could actually work? The answer, I believe, lies not in replacing attorneys but in automating their invisible infrastructure. Imagine a system that understands an unstructured email from a local patent attorney in Japan, extracts key details – deadlines, tasks, fees – and structures them automatically in your own database. Imagine it replying in a format that the Japanese attorney immediately understands, without anyone manually rewriting or forwarding.

This is not science fiction; it’s what narrow, specialised AI agents – vertical AI – can already begin to do. Instead of acting as another intermediary, they function as an intelligent layer that translates, organises, and executes. They preserve direct relationships while removing the human ‘busywork’.

Over the past few years, I’ve been part of a team testing these ideas in real-world conditions, integrating AI into global IP workflows. That work culminated in the creation of iPNOTE, a platform built to streamline foreign filings by combining direct access to verified local IP attorneys worldwide with AI-powered operational automation. The concept is simple: keep the people who matter, but eliminate the friction that doesn’t.

After more than a decade inside this problem – first as an attorney, then as someone trying to solve it from within – I’ve learnt that technology cannot change the nature of legal expertise, but it can transform its mechanics. When it does, the old trade-off between time and money in global IP management will finally disappear – replaced by something far more valuable: clarity.