A global fund manager is of the belief market sentiment expressing fear over the disruption artificial intelligence (AI) is causing in relation to software businesses is overblown.
According to Schroders head of investment directors for public and private markets Claire Smith, the market has overreacted in this sense.
“There’s been a view that AI is killing software. AI is absolutely reshaping parts of the software market, but the idea that every software company is suddenly at risk simply isn’t how private investors are thinking about it,” Smith indicated.
“When you look underneath the surface, many of these businesses still have highly sticky customer bases, proprietary data and critical functionality.”
To test this conclusion, Schroders performed an “AI threat analysis matrix” to identify the businesses that faced genuine disruption risk and which were likely to remain resilient based on the investments it holds in software companies.
“We assessed whether AI could reduce the number of software seats being sold or potentially make a platform redundant altogether,” Smith revealed.
“In our semi-liquid private equity fund, only around 2 per cent of the portfolio fell into what we classified as high risk.”
She pointed out it was still difficult to replace software businesses servicing highly specialised industries, such as those handling sensitive or operationally critical data.
“You’re not going to vibe-code your way around payroll systems handling confidential patient data. Businesses still need reliability, compliance and security. AI is not eliminating that,” she suggested.
However, she acknowledged AI disruption is a real risk for some companies, but at the same time this fact does not disparage the investment opportunities the technological developments are presenting.
“We have invested in AI-linked businesses including a data annotation company servicing major artificial intelligence groups including OpenAI, Meta and Nvidia,” she confirmed.
“We prefer businesses that are benefiting from the growth in AI infrastructure, rather than trying to predict which individual AI applications will ultimately win.”