The Economics of Research: Why Faster Is No Longer Enough

AI is rapidly reducing the cost of finding and processing information. But the economics of IP and innovation research should not be judged by speed and cost alone. As Justin Delfino, Partner at S&W ClearViewIP, explains in this article, the more important question is whether research improves the decisions that determine where organisations invest, protect, partner and stop.

I have spent more than 20 years working in IP, R&D and innovation research. During that time, the economics of research have often been framed around familiar measures: hours, people, turnaround time, output and cost.

There is nothing wrong with those measures. I have spent a significant part of my career working with them. But they measure the economics of producing research rather than the economics of the decision that research supports.

Those are not the same thing.

Consider a freedom-to-operate study supporting a major product launch. The economically important question is unlikely to be whether the research cost EUR 10,000 or EUR 8,000. It is whether the work gave the business sufficient confidence to invest, identified a risk before capital was committed, or provided the evidence required to change course.

The same applies to an innovation landscape, technology-scouting exercise, prior-art search or portfolio review.

The report is an output. The decision is the outcome.

In July 2026, WIPO and Luiss Business School reported that intangible investment across the 29 economies covered by their analysis exceeded USD 10 trillion in 2025. Those economies represent about 57% of world GDP. In real terms, intangible investment grew at a compound annual rate of 3.5% between 2008 and 2025 - 3.6 times the rate of tangible investment.

The category is broader than patents. It includes R&D, software and data, design, brands and organisational know-how. That is precisely why the trend matters to the IP profession: many of the questions crossing an IP team's desk sit inside much larger decisions about innovation and capital allocation.

Should we continue developing this technology? Where should the next R&D investment go? Does the evidence support filing? Are we entering a crowded technology space? Should we build, buy, partner or license? Does the portfolio contain assets with commercial value that we have overlooked? Should this programme continue at all?

These are not simply research questions. They are investment decisions involving intangible assets.

And that changes how I think research should be designed.

One of the consequences of better technology is that producing information is becoming cheaper and faster.

That is overwhelmingly positive. But more information does not automatically create more insight.

A patent landscape containing thousands of families may be technically excellent and commercially unhelpful. A well-executed prior-art search can still fail to answer the question the business really needed answered. An AI system can process an enormous body of information without possessing an important piece of internal context held by an engineer who has worked on the technology for fifteen years.

The scarce resource is changing.

Historically, much of the cost of research came from finding and processing information. Increasingly, I think the scarce resource will be judgement.

Which evidence matters? What is missing? Where do different sources disagree? What does an apparently strong signal actually mean for this particular business? And ultimately: so what?

That last question is deceptively difficult.

Some years ago, I worked on a project examining an emerging-market opportunity for a major global consumer business.

Patent research alone could not answer the question. We combined patent information with scientific and technical literature, market evidence and primary research. The interesting part was not any individual dataset. It was what happened when those different forms of evidence were brought together and experienced researchers started forming a view.

That work ultimately contributed to a new consumer product being launched in an Asian market.

I cannot remember how many patents we reviewed. I remember the decision.

More recently, I have been involved in redesigning the process through which inventions move towards decisions inside a large innovation-led organisation.

Different problem. Same lesson.

The economic opportunity was not simply to make each individual research activity cheaper. It was to improve the quality, consistency and timing of decisions being made repeatedly across an innovation portfolio.

A small improvement in one decision may not matter very much. A small improvement repeated hundreds of times can matter enormously.

AI should materially reduce the effort required for some research activities. It should increase speed. It should allow experts to examine more information and identify relationships that might previously have been difficult to see.

All of that is valuable.

But I do not believe the end point is simply replacing expensive research with cheaper automated research. That would be an extraordinarily narrow use of the technology.

The bigger opportunity is to change the allocation of effort.

If machines can increasingly undertake repetitive work involved in finding, organising and processing information, experts can spend proportionately more time interpreting it, challenging it and connecting it to the commercial question.

The economic model therefore starts to change from:

How cheaply can we produce the research?

to:

How effectively can we combine technology, evidence and expertise to improve the decision?

That is a much more interesting proposition.

There is another consequence. Many important innovation decisions do not sit neatly inside an IP department.

A decision about a technology can simultaneously affect R&D expenditure, patent strategy, tax incentives, future capital requirements, licensing opportunities, M&A and ultimately enterprise value.

Yet organisations understandably structure those disciplines separately. IP has its data. R&D has its technical knowledge. Finance has its investment model. Tax has its evidence requirements. Corporate development has its transaction perspective.

Each can be excellent individually and the overall decision can still be suboptimal.

I do not think the answer is to collapse those disciplines together.

I do think the research supporting major innovation decisions increasingly needs to connect them.

That requires broader evidence, better technology and, importantly, people with enough experience to form an opinion rather than simply deliver an output.

This thinking has influenced what we are building at S&W ClearViewIP.

As we expand our Research & Analytical Services capability, the ambition is not to recreate a traditional research outsourcing model with newer technology. It is to combine scalable research capability, modern technology and experienced IP, R&D and innovation professionals with the wider advisory perspective needed to understand the commercial question.

Search still matters enormously. Quality matters even more.

But the objective is not research for research's sake. It is to get from information to insight, and from insight to a better decision, as efficiently and consistently as possible.

That means knowing when technology should do the work, when an expert needs to do it, and when the real answer requires IP research to be connected with a wider commercial, technical, tax, financial or transaction perspective.

The 2026 IP Service World agenda shows just how quickly the industry is moving: agentic AI, automated research, AI-assisted patent workflows, decision platforms and new approaches to search feature prominently across the programme.

Those developments matter. They will change our profession.

But perhaps we should be careful about what we choose to optimise.

If research becomes twice as fast but does not improve the decision, we have certainly improved productivity. I am less certain that we have created much additional value.

For the next generation of IP and innovation research, I think the more important question will be:

Did we bring the right evidence, technology and expertise together, at the right point, to make a better decision?

That is harder to measure than hours, turnaround time or cost per search.

It may also be considerably more valuable.