Software has been a matter of debate in public markets this year, as investors weigh whether AI represents disruption or opportunity for the sector.

In this conversation Hg's Nic Humphries, speaks with Jim Strang, Chairman of HgT, the publicly-traded access point to Hg, about how we're thinking about AI across the portfolio - from the internal data science team built over a decade ago, to Hg Catalyst's 100+ AI engineers embedded in portfolio companies, to the early evidence of margin expansion showing up in the numbers.

For those short on time, the key takeaways are below.

Key takeaways

AI is augmentation, not substitution - for the right kind of software.

Nic argues the market's caution is rational for generalist investors who can't distinguish which software businesses are exposed to disruption and which stand to benefit. Hg's view: businesses with deep customer trust and compliance-critical workflows (he uses vertical software for small business back-office functions as the example) are seeing AI lower their development costs and expand their addressable market, not shrink it.

A decade of AI infrastructure, not a reaction to ChatGPT.

Hg's AI programme started over ten years ago as a data science function analysing portfolio metrics, well before generative AI entered the public conversation. That history compounds into two capabilities today: an internal team of 20+ data scientists working across the portfolio, and Hg Catalyst - Hg's own AI engineering function, now over 100 people - which embeds with portfolio company development teams for roughly six to nine months to build and launch AI products before handing over to the internal team.

The economics are already showing up.

Some companies in the Hg family now have $30–50m of annual recurring revenue from AI products growing 50–100% year-on-year but, more broadly, Nic points to margin expansion of 3–5 percentage points a year across the portfolio. This is well above the roughly 1 point historically typical and is driven by AI productivity gains in engineering and customer support, without needing to grow headcount.

AI leadership is now a core investment criterion.

Looking at new investments, Hg is applying a simple filter: is this business the AI leader in its niche? That means either backing native AI companies reaching a scale Hg can invest in, or established software businesses that have already launched AI products with real customer traction.

Proof points from recent exits.

Nic cites two December exits where AI product launches drove premiums of roughly 30–40% above what the underlying business would otherwise have commanded, as strategic acquirers moved early to avoid paying a much larger premium later.

Partnerships and talent as a compounding advantage.

Early relationships with AI labs and tools (Anthropic among them) gave Hg's portfolio companies visibility into new technology ahead of the market, which Nic credits with helping Hg retain specialist AI talent that might otherwise be drawn to start-ups.