Franklin Templeton on where AI is creating winners and casualties

Enterprise software is going through its biggest reset in a generation. The question is not whether AI will disrupt the sector. It already has. The more useful question for advisers and asset managers is where value is being created and where it is quietly eroding.

Franklin Templeton’s Jonathan Curtis, portfolio manager at Franklin Equity, has a clear framework for thinking through exactly that. He offers a modernized lens for enterprise software investment in the age of artificial intelligence.

The two forces reshaping the sector

Two separate but related forces are at work, Curtis says.

The first is the collapse in the cost of creating software. Since 2022, AI has moved from a novelty to a genuine engineering partner. By late 2025, leading AI coding tools had nearly matched the output of skilled human developers, and they are continuing to improve. The practical consequence is that the volume of software written and deployed globally will increase dramatically. More software, built faster and at lower cost.

The second disruption is broader and more uncomfortable for many software vendors. The same AI systems that write code can now draft documents, synthesise research, manage workflows and automate complex analysis. Enterprises are realising that AI can make entire organisations leaner.

“That realisation raises an uncomfortable question for many software vendors,” Curtis says. “If AI can perform the work, how many human software seats does an enterprise customer need?”

That question is already showing up in earnings results, renewal conversations and customer behaviour across the sector, fundamentally shifting the risk parameters of investing in enterprise software.

Three camps, three very different outlooks

Curtis argues that the most important insight from the recent earnings season is that software is no longer a single investment narrative. The sector has split into three distinct camps, and treating them as one asset class is where investors are likely to make costly mistakes.

The first camp covers companies managing the growing complexity of enterprise AI deployment, including monitoring, security, governance and orchestration. These businesses are experiencing accelerating demand. “AI is the engine driving their growth, not a threat to it,” Curtis says.

The second camp is where the pain is most visible. These are companies built on the assumption that enterprise headcount would grow predictably and that each employee would need a licence. They are not all in decline, but they face a difficult transition at precisely the moment customers are questioning how much value they actually need from a seat-based model.

The third camp is where Curtis sees the most compelling long-term opportunity for enterprise software investment. These are companies whose products serve as the connective tissue of enterprise AI deployment: workflow orchestration, process automation and systems of record with critical enterprise context. “Their value proposition is strengthening as AI complexity grows, and the best of them are growing faster than their pre-AI trajectory with pricing power intact,” he says.

What to look for in a winner

For those looking to refine their enterprise software investment strategy, Curtis identifies five characteristics that separate the well-positioned from the vulnerable.

Deep integration that creates genuine switching costs. AI-driven revenue acceleration traceable through usage growth. Stable or improving gross margins. Effective use of AI in their own internal operations. And consumption or outcome-based monetisation models, rather than flat per-seat pricing.

That last point matters more than it might seem. Companies that have shifted to usage or outcome-based pricing are aligning their revenue model with how AI actually creates value for customers. Flat seat licences, by contrast, become harder to justify as headcount falls and AI takes on more of the work.

What this means for advisers and asset managers

The broader message from Curtis is a direct challenge to the way software has been categorised and evaluated as an investment.

“AI is reallocating value within software, creating clear winners and exposing vulnerabilities in business models that have worked well for the past two decades. We believe investors who treat software as a uniform asset class will make costly mistakes.”

For advisers with client exposure to technology funds or global equity strategies with significant software weightings, the composition of those holdings matters more than it did two years ago.

A fund with heavy exposure to seat-based licence businesses faces a very different outlook from one concentrated in orchestration, governance and AI infrastructure plays.

The total volume of software consumed globally will almost certainly grow, as Curtis notes. But the companies capturing that growth will not look like the ones that led the last decade. Knowing the difference is where the work starts.