AI Trends in Enterprise Software M&A
The trends keep trending.
What’s the state of middle-market enterprise software, AI, and capital markets as we look toward the last 5 months of 2026?
Let’s get into it.
Very few enterprise software firms have any designers on staff, which means user experience is an afterthought (e.g., I pay you, just deal with it, or plug it into your AI chat tool). This is a sector-wide opportunity.
Many have 95%+ of new code written by AI (though humans still do PR reviews).
Buy-side investors are looking at the AI story before financials, which seems crazy at first blush, but makes sense based on the sheer impact the technology is having across all industries.
AI disruption risk is on everyone’s mind and has been a hot offering for us.
Durable moats tend to center around compliance and regulatory lock-in, software used for managing deep operational workflows, and some elements of proprietary data.
AI use cases are in production and beginning to generate recurring revenue.
Customers can now DIY many things themselves so simple products are getting the boot (e.g., basic survey tools, CRMs, etc), though large, complex businesses with hardware installations are less exposed.
AI-native startups are disrupting incumbents at the interface layer first (see our point above on no UX designers on staff), and the logic layer second (see our below point on no proprietary models or AI folks on staff). This has the knock-on impact of compressing profit margins if you don’t defend appropriately.
Few have large AI/ML or Data Science teams on staff (e.g., less than 5), other than AdTech (algo matching viewer with publisher) or FinTech (underwriting or risk models).
Most use OpenAI or Anthropic models in the SDLC process and in their products.
While many may have aspects of proprietary data, almost none have proprietary models. This is another sector-wide opportunity.
The transition of existing enterprise software products to AI-native, end-to-end platforms is more rare. We’ve seen it when a company has a true innovation team on staff and the competitive pressure/vertical investment is high, but this represents early indications of where things are headed in the medium term.
There’s a wide variety in the way customers perceive AI in enterprise software, depending on the industry vertical. Some actively consider it as key purchase criteria while others are more focused on their own business outcomes and don’t care “how the sausage is made”.
Some novel product thinkers are beginning to allow customers to edit software themselves (this is nascent, but can see this taking off to remove the DIY risk and create higher lock-in/personalization).
On the sophistication continuum of Data Readiness to Insight/Chat to Agentic Action to Autonomous Action, most are somewhere between AI Chat and AI Agents, though the more advanced product development teams are starting to play with the concept of invisible software (e.g., suggested task completions presented to users for acceptance, which front-runs full autonomy).
Rule of 40 (ARR Growth % + EBITDA Margin %) still matters, with most companies going to the capital markets somewhere between 40% and 70%. Palantir, at its size, is posting insane numbers of 155%, which is among the highest ever reported by a public software company. About 2/3rds of that is revenue growth.
Customer retention, revenue retention, and churn are still big deals, especially when capital providers are looking for early indications that AI may be disrupting cash flows.
Some cash cow businesses, with low or no growth, still find buyers as part of a carve out or tuck-in simply because it comes with profit, product, and growth potential.
We’ve even seen businesses that were declining top-line, but get scooped up because the valuation/price was low enough that it made sense for the buy-side.
Sell-side prep continues to be a hot offering because boards and management teams don’t want to walk into an intense and fast buy-side process unprepared for how deals are viewed today. This is coming up more often so teams can address short-term issues before they go to market.
M&A activity in the space is expected to be strong for the rest of 2026, especially for good businesses with strong financials and product.
AI is not going away: you need to have strong use cases driving ARR growth with proper governance, MLOps, a solid team, and defensibility built in.
We’ve heard from multiple investment committees that they do not want to inherit a mess or buy something that’s going to make them look bad in 6 months or a year because of the changing market and product dynamics (e.g., is your architecture and infrastructure a mess, do you understand how to productize AI to drive growth, are you using AI appropriately in the SDLC process, how are competitors and customers responding to your company and product, etc).
Overall, the SaaSpocalypse has been overstated. These businesses are not being destroyed overnight.
There is nuance involved in every deal and every business. That said, if you’re sitting on your hands and doing nothing, then you’re at risk. We’ve only seen a few management teams say “AI doesn’t impact us”, but as each day clicks by, it’s likely that sentiment goes away entirely.
The prepared teams who take AI seriously, and are at the front of the pack with AI-enabled products that are driving revenue are in good shape. We expect that trend to continue over time. As a result, the art and science of product management matters more and more, especially as these folks fight to invest in next-gen user experiences and proprietary models that make customer’s lives easier and outcomes faster to achieve.
That’s where we see the next 12 months headed as agentic coding tools eat product development teams and the entire org from the inside out.
Open-source software and model routers help, but even your own in-house built agents need appropriate harnesses to meet the bar of “enterprise-grade”. Secure, permissioned, reliable, and governed against drift and out-of-control token spends.
We see business models changing with commercial teams continuing to experiment, but the end-game is likely somewhere between enterprise software companies eating variable token costs while offering stable subscription prices for customers, with your product add-ons, and overages for things that get way out of control (but they shouldn’t if you’re doing product correctly). Essentially, the enterprise software companies “eat” the messy-middle variable changes in tokens, but spread it out over more customers so they can plan. Stay within the bounding boxes and you’re good.
Have a great back-to-school season and don’t forget to buy your new Apple devices 🙂
—Sean
