Writing a patent opinion was never the hard part for Dimitris Giannoccaro, IamIP’s CEO. Finding the right documents was — and as a former patent examiner working two-day application cycles, he spent far more of that time searching than judging.
An entire day could disappear into search, sometimes more than one. You start from a technical problem, try a handful of keywords, follow the references those turn up, check nearby classifications, look at what competitors have filed, then adjust and start again from a different angle. You don’t know what you’ve missed until you’ve already looked in three or four other places.
Search is its own skill, separate from patent expertise. Plenty of examiners were sharp on substance and only average at finding material, and the reverse was just as common. That divide has shaped the profession for decades and cost the industry more time than most people account for.
Two Skills, One Bottleneck
That gap between finding and judging is finally closing.
AI search tools are strong enough now to take on the searching. Not the judgment. The searching. AI isn’t replacing patent professionals, and it isn’t deciding whether something is patentable, infringing, or in scope. What it’s changing is how much of a professional’s day goes to finding material before any of that judgment can start.
Faster search tools wouldn’t have made examiners think faster. They would have left a lot less time spent hunting for material, and a lot more time on the part of the job that actually required expertise.
Where the Time Really Goes
Analysis gets better with experience. The more patents someone has reviewed, the sharper their instincts become. Search doesn’t work the same way. The volume of material to search through keeps growing regardless of how experienced you are. WIPO’s most recent figures put the number of international search reports issued worldwide at roughly 264,000 in a single year, with the EPO alone responsible for nearly 82,000 of them. Every one of those reports starts the same way an examiner’s day always did: with someone searching before anyone can judge. One skill compounds with time. The other just gets harder.
That gap costs more than hours. When search eats most of a working day, the final opinion still gets written. What suffers is everything that should have happened before it: fewer angles tried, fewer adjacent classifications checked, fewer competitor filings cross-referenced. The output still looks complete. It isn’t always as thorough as it should have been. The constraint was never expertise. It was the time expertise had to compete with.
Who Gets to Do Good IP Work
There’s a second shift worth naming, and it has more to do with access than time.
Boolean search strategy has always been its own specialism. For years, search quality depended almost as much on who was running the query as on who understood the subject matter. A strong engineer with a weak search strategy could miss what an average engineer with a good one would find. When someone can describe a technical problem in plain language and still get relevant results back, that gap narrows. It changes who can do solid IP work in the first place, not just how quickly they do it.
It’s also why the tools that help here tend to be narrow rather than general. A general-purpose AI model doesn’t know what a classification system is, why legal status matters, or how claim language is structured. The tools worth using are trained specifically on patent data and patent problems, because that’s where the understanding has to live.
Insight Over Speed
The teams getting this right aren’t the ones with the fastest search box. They’re the ones who treat search and judgment as two different problems and give each the right kind of support. Speed on its own was never the goal. Redirecting time toward judgment was.
That’s the real shift underway in patent work: not AI making the calls, but AI making sure people spend their time on the calls that matter, instead of on the search work that has to happen before those calls can be made at all.
Insights by Dimitris Giannoccaro
