AI isn't killing software. It's separating the winners from the rest.

The year has opened on a challenging note for publicly listed software companies. At time of publishing, the S&P Software Index is still down roughly 14%, reflecting a broad sell off across the sector. Volatile markets, amplified by dramatic media headlines and AI companies fighting for attention, have only added to investor unease. If you hold investments in technology, it’s natural to feel uneasy. We understand why. But we also think the public market has reacted indiscriminately. It’s treating all software as if it were the same. It isn’t.

Two very different types of software

Think about the software you encounter at work. Some of it is broad and general and can be used by any business in any sector – which is why it’s often called ‘horizontal’. By this I mean basic productivity tools, generic customer databases, simple analytics dashboards. These products are designed to organise information for people to look at and use more easily. But when AI can talk directly to this same data, within a system or ‘interface’ that you can design yourself, that visual layer at the front end becomes less valuable. We think the market is right to question the future of those businesses. Then there’s a different category entirely. Specialist ‘vertical’ software that manages mission-critical, complex processes in specific, often regulated, industries. Think tax filings, payroll, legal, compliance, healthcare, regulatory reporting. These systems don’t just display data. They execute complicated workflows, handle sensitive information, need to be right every time, and often file information directly with authorities or support care for a patient. A wrong answer isn’t an inconvenience; it could become a legal liability and lead to broader implications. This is the kind of software we invest in at Hg, and it has four characteristics that provide real advantages when building agentic AI products on top. We call them the ‘four Ds’:

• Deep proprietary data built up over years

• Domain knowledge embedded into the product over decades

• Distribution through trusted customer relationships and a network built up over time

• Deterministic outcomes (meaning these processes require specificity and accuracy, not approximations)

AI can make these businesses more valuable

Here’s what the broad sell-off misses. For specialist software companies who move quickly to leverage these advantages, we anticipate AI will be an accelerant, not a threat. It will enable them to evolve from systems that store and organise information, into systems that actually do the work, handling tasks end-to-end under human oversight.

Imagine software that you may use for work. Much of it still requires some form of repetitive (and tedious!) action, perhaps data entry, or analysing the output of one system to share elsewhere. Now imagine an AI ‘agent’ doing these actions for you, with your oversight and approval. This saves you time, increases productivity and therefore holds intrinsic value.

That’s a significant shift. The global software market today is worth roughly $1 trillion(1). But when software starts completing human tasks rather than just supporting them, it begins to compete for a share of the $50 trillion-plus the world spends on human labour each year(2). Importantly, we are also seeing that margins on these new AI ‘products of action’ are comparable to traditional SaaS margins, only they can now deliver much more value to their customers. That’s why we believe the companies that move quickly could be worth considerably more, not less.

What we’re doing about it

We’ve spent the best part of a decade building what we believe is one of the most comprehensive AI capabilities of any private equity firm, precisely because we saw this moment coming.

One example includes Hg Catalyst, our dedicated AI product incubator. It allows us to deploy around 100 engineers, product managers, and designers directly inside our portfolio companies to rapidly accelerate the build of agentic AI products. When a breakthrough happens at one company, we can roll it across almost 60 businesses within weeks. That compounding effect is powerful and hard to replicate.

The early results have been striking too. As of March 2026, we have more than 1,600 AI projects live across our portfolio and have launched over 100 AI-powered products, delivering real revenue and a large portion of new customer bookings. Together, this has already delivered a positive impact to the bottom-line of the relevant companies that’s growing rapidly.

Hg Catalyst in action: GTreasury

GTreasury is a treasury and risk management platform used by corporate finance teams worldwide to manage cash, forecast liquidity, and control risk exposure. It's a good example of the kind of specialist, mission-critical software we've been describing, and of what happens when you combine that foundation with dedicated AI capability.

Hg Catalyst kicked off with a two-day hackathon, mapping GTreasury's core workflows and producing three working prototypes. The winning concept went from demo to live, priced, packaged product in just eight weeks. Two further AI features followed within 12 weeks, and the company compressed its release cycles from monthly to weekly.

The collaboration produced ‘GSmart AI’, GTreasury's new AI-powered forecasting and insights product. The impact was apparent almost immediately. Within weeks of launch, the product was generating significant new revenue and a strong pipeline. Around 40% of all new inbound customer enquiries were driven by GSmart, and the product hit 100% weekly active usage from launch. Importantly, GTreasury's AI strategy also became a critical factor in Ripple's subsequent decision to acquire the business.

GTreasury is not an isolated example. Ripple's acquisition of the business at a 95%+ uplift was in large part driven by its AI transformation. Intelerad, a leading provider of medical imaging software, was similarly acquired by GE Healthcare in 2025 at a significant uplift, following AI transformation of its diagnostic imaging platform.Together, these exits reflect broader portfolio momentum. Our companies continue to deliver strong operating performance, growing revenues at +20% with 30%+ EBITDA margins.(3)

Why this moment matters for investors

When the market reprices an entire sector at once, it usually doesn’t distinguish between businesses that are genuinely at risk and those that are temporarily caught up in the uncertainty. For a generalist, that’s uncomfortable. For a specialist with over two decades investing in technology and tech-enabled services, it’s exactly the environment we were built for. We anticipate it will create some great opportunities for those who know what to look for.

To succeed, we think you need to identify which workflows are defensible, which data advantages are real, and which management teams can lead an AI-first transformation. That’s the work we’ve been doing for over two decades – meticulously studying workflows, determining what users really need most and what they will really find valuable as customer. Combine this with our ability to invest in innovation, fundamentally transforming the products and operations of the businesses we partner with, and you have a compelling recipe for long-term returns.

In short, the narrative that AI kills all software is too blunt. The reality is much more interesting. AI will create clear winners for the future and we see those winners as those with proprietary data, deep expertise, and the ability to become AI-first. That’s where we will continue to focus.

To save or share this article, find a PDF downloadable version here

Notes:

Information is at March 2026 unless stated otherwise. Hg’s views and opinions are provided for general informational and discussion purposes only and should not be construed as investment, legal, accounting, or other professional advice. Case studies presented herein are for illustrative purposes only and are intended to illustrate Hg’s sourcing experience and the profile and types of investments previously pursued by Hg. It should not be assumed that investments made in the future will be comparable in quality or performance to the investments described herein. Further, references to the investments included in the illustrative case studies should not be construed as a recommendation of any particular investment or security. Statements contained in this letter are based on current expectations or estimates and are subject to a number of risks and uncertainties. Actual results, performance, prospects or opportunities could differ materially from those expressed in or implied by these statements and you should not place any undue reliance on these statements. Past performance is not a reliable indicator of future performance.

Footnotes:

1 Gartner (Feb 3, 2026) projects total global **software spending will be more than $1.4 trillion in 2026 (a major segment of the overall IT spend forecast).

2 International Labour Organization, Global Wage Report 2024-25 and ILOSTAT labour income data, provide the basis for global total labour compensation estimates measured in the tens of trillions of U.S. dollars annually (often cited around USD 50 trillion).

3 Based on Hg portfolio-level metrics for LTM September 2025, adjusted for post-period events. Weighted average across Hg's active software portfolio companies. Revenue growth +23% (+12% organic); EBITDA growth +27% (+18% organic). Individual company performance varies. Past performance is not indicative of future results.