In early June 2026, Hg, one of Europe's largest software investors, published a remarkably clear piece: A new lens: Investing in an AI world. Three partners describe how they look at software companies now that AI is changing the rules. This is not a marketing story. It is the investment logic of a firm that brings agentic AI into practice every day across dozens of software companies. The reasoning starts from a conviction Hg has held for two decades. Technology raises the productivity of professionals, and the companies that best absorb a new capability into their workflows build durable customer value. From on-premise to cloud, the same pattern every time. AI is the next wave. Software is no longer limited to recording and executing; it analyses, reasons and acts, with agents doing part of the work themselves. As a result, the addressable market grows substantially. Hg calls this the new "agentic TAM". The question that follows is the right one: who has the best foundation to capture that agentic TAM? Hg now weighs four criteria more heavily than ever.
Four criteria for the winners
One. World-class, product-led leadership. The leaders rethink what is possible with AI. They don't optimise a tired product, they build anew. The strongest AI strategies in Hg's portfolio come from teams that know their customers' workflows in detail and build AI seamlessly into them, so adoption happens naturally.
Two. Deep product logic in verticalised use cases. Every technology business encodes domain knowledge into its product: rules, processes, regulatory frameworks, accumulated experience. Where that logic is deep and constantly changing, it is almost impossible to replicate. Think of platforms that safeguard jurisdictional rules, consolidation rules and exceptions across hundreds of entities, at one hundred percent accuracy. Companies with deep vertical product logic have reason to be glad every time the underlying models get better.
Three. Critical outcomes where the cost of failure is high. In many of these sectors, the output is the outcome itself: a filed annual report, a payroll run, a compliance certificate, an audit. If it goes wrong, the consequences are immediate and often irreversible. Customers choose their software for the assurance that it will be correct. AI that meets the accuracy bar lets the professional do more, and so raises revenue per customer.
Four. Proprietary data you create, not assemble. Public information that once required special tools is now available in seconds and has lost its economic value. What counts is a data flywheel from platform usage: transactions, operational logs, decision histories. Data with context, which cannot be reconstructed elsewhere and makes the product smarter every month.
Two observations complete the picture. AI adoption does not happen everywhere at once: sectors with longer adoption horizons, high costs of failure and a lot of proprietary data favour precisely the established players with a product-led team at the helm. And, importantly for those building from Europe: Europe's regulatory fragmentation, often seen as a burden, is a structural advantage for these companies. AI makes that advantage bigger.
Hg backs it with figures from its own portfolio: more than a hundred AI products launched, build times shrinking from nine months to under three, and a Value Creation team of more than 150 AI professionals. The conclusion is sober. Not every software company survives the AI wave. Speed without foundation is fragile.
Where Fyrm.ai stands in this picture
Anyone who reads this piece from Fyrm.ai's vantage point recognises the profile. The criteria Hg describes almost coincide with the reason we were founded.
Our DNA is technical. AI and software engineering, computer science and security, combined with a thorough understanding of business processes. That is exactly the product-led foundation Hg names as its first condition. And where most organisations quickly build a prototype and get stuck in the proof-of-concept, we build on to a Full Product AES: an Agentic AI Enterprise System, integrated into the existing business systems, with all security layers built in. Our vision can be summed up in four words: See it. Scale it.
We deliberately chose highly regulated markets: Telecom, financial and legal services, and Logistics. It is precisely there that legally established, standardised rules apply. That makes an AES modular and scalable, periodically updatable as ongoing SaaS, and built on deep domain knowledge of end-to-end processes that competitors find hard to enter. It is exactly the deep product logic Hg looks for, with long-term customer relationships and high switching costs as a result.
In these markets, the outcome is the product. A correct processing run, a watertight compliance filing, a process that keeps running. The cost of failure is high, which is why governance and security sit in our foundation rather than as an afterthought. And because an AES runs inside our customers' core processes, as a long-term SaaS partner, that is also where the operational data with context arises that keeps making a product smarter. That is the fourth building block Hg names.
At the front end, we stay sharp. We are agnostic across all major frontier and cloud platforms, databases and business systems, so every improvement in the underlying models makes our product stronger rather than more vulnerable. And through partnerships with selected universities we educate, offer internships and recruit top talent, so the team building this stays ahead.
Finally, the European point. We build from Rotterdam, for highly regulated markets in the Netherlands and Europe. The regulatory fragmentation Hg identifies as a structural advantage is the playing field we are set up for.
Hg describes the winners of the AI world as companies with product-led leadership, deep domain logic, accountability for critical outcomes, and data that compounds. We are built on that profile. Agentic AI will become the new fundamental layer for the decades ahead, just as the internet did from the 1990s.
Fyrm.ai. Orchestrated Intelligence. See it. Scale it.
Source: Hg, A new lens: Investing in an AI world (10 June 2026), by Benedikt Joeris, Jonathan Wulkan and Joris Van Gool, partners at Hg.
