Cor­por­ate in­nov­a­tion has stalled – and AI risks deep­en­ing the chal­lenge

Once a month, researchers at Copenhagen Business School provide Børsen readers with a current and research-based perspective on the challenges facing leaders.

This time professor, Giacomo Marchesini points out that AI is effective at utilising the knowledge we already have – but not good at making entirely new discoveries.
Giacomo Marchesini, Assistant Professor at the Department of Strategy and Innovation at Copenhagen Business School.

Investments in research and development have increased for decades, and Danish companies now spend more on innovation than firms in most other countries. On paper, we should be in a golden age of discovery.

Yet major breakthroughs – the ones that cure diseases, create new industries or radically improve productivity – are becoming rarer. And artificial intelligence may well worsen the trend rather than reverse it.

This is not only a research and development issue. It is a leadership challenge. The decisions leaders make about incentives, time horizons, risk appetite and the use of AI increasingly determine whether innovation systems lead to genuine breakthroughs or simply produce more of the same. The popular “fail fast” approach may not be as effective as it seems.

We invest more but discover less

Across industries, we see a consistent decline in novelty. Companies generate more patents than ever, but the share that is genuinely new or disruptive continues to fall.

In the pharmaceutical industry, the costs and effort required to develop a new drug have risen sharply in spite of strong scientific tools and access to unprecedented amounts of data.

As knowledge accumulates, innovation naturally becomes more complex, but we also risk innovating more without innovating better.

The explanation is not how much companies spend on research and development but how they make decisions about innovation.

A good example comes from my research on how firms respond to short-term financial pressure. We found that pressure to improve short-term performance affects the types of innovation projects companies choose to pursue.

In a study of European firms subject to mandatory quarterly reporting, I found that innovative companies did not cut their R&D budgets – some even increased them as capital became cheaper.

But the nature of their innovation shifted. They moved towards safer, more incremental projects and focused on technologies they already knew. They avoided new or riskier areas. Exploration declined even as overall R&D increased.

If corporate innovation is stalling, leaders cannot remain passive. More funding alone will not solve the problem. What matters is rethinking the habits, decision processes and incentives that determine whether good ideas are allowed to grow – or disappear.

AI is not the solution (at least not yet)

Many hope artificial intelligence can be the answer: better data, faster analysis, smarter pattern recognition. And AI adoption is spreading quickly in Danish firms, almost twice as fast as the EU average. But new research suggests that most AI projects do not deliver the promised value, and only a small fraction lead to breakthroughs.

AI does not appear to address the fundamental challenge of breakthrough innovation. A simple reason is that AI is trained on existing information. It amplifies what we already know rather than guiding us to areas we have not explored or helping us find ideas that do not yet exist. In fact, AI can reinforce the “streetlight effect” – we look where the light is strongest, not where the next breakthrough may be hiding.

AI excels at exploitation – refining, optimising and accelerating – but breakthroughs happen in exploration through discovery and experimentation. And exploration requires something AI does not yet have: curiosity, tolerance for being wrong and a willingness to follow faint early signals. This is where leaders play a vital role.

Three ways leaders can strengthen innovation without waiting for AI

1. Update your innovation expertise – scientifically.
Leaders make countless decisions every year, yet many rely on intuition or habits within the industry. Innovation research has advanced significantly, but only a fraction reaches companies. Many leaders still rely on “best practices” that lag behind the latest evidence – and some can even be misleading. This knowledge gap is costly. When leaders misunderstand how new knowledge emerges, they design systems that inadvertently hinder innovation. So: invest in evidence-based innovation training, conduct proper innovation audits and use metrics that capture novelty and long-term impact.

2. Focus on “solutions looking for problems.”
Governments, including the Danish government, and companies are investing heavily in problem-oriented innovation to tackle climate and health challenges. This can unintentionally narrow the space for discovery in the exploration process. Cognitive research shows that when problems are defined too early and resources are tied to them, we reduce the chances of unexpected discoveries. Historically, science often works in the opposite direction: penicillin, X-rays and the microwave oven did not begin as solutions to predefined problems. They started with surprising observations and the question: “That is strange… what could we use it for?” Leaders should create room for this kind of serendipity. So: allocate modest budgets for open exploration, allow teams to follow unexpected results, avoid KPIs that confine creativity and do not require every idea to be justified in advance.

3. Rethink the “fail fast” doctrine.
“Fail fast, fail often” assumes that all failures are alike and should be abandoned quickly. They are not. Many breakthrough projects fail repeatedly before they succeed. In ongoing research on drug development, we see that the most successful companies do not give up early. They build broad portfolios of approaches around the same idea, and when one fails, they shift resources to new alternatives closely related to the failed approach – not to something entirely different. This helps them diagnose and learn from setbacks. Firms that rely solely on “fail fast” often learn nothing and waste resources, much like throwing ideas at a wall and hoping some will stick.

AI will not save us from poor innovation habits

AI will transform large parts of the business community, but it cannot compensate for outdated leadership practices, narrow exploration processes or skewed incentives. Breakthrough innovation requires leaders who understand innovation research, protect open spaces for exploration and make decisions rooted in evidence rather than fashion.

What firms need most is intelligent leadership – before artificial intelligence.