Most commentary on enterprise AI adoption sits at one of two extremes. Either macro forecasts about trillions of dollars of value, or anecdotes about a single tool deployment. What's harder to find is the middle ground: what's actually driving buying decisions right now, and how that translates into revenue for the companies catching the wave.

We sit in a useful position to observe this. Three Haatch portfolio companies posted standout Q1 2026 results, and the more interesting story is that each is succeeding for a different reason. Here's what each tells us about how AI adoption is reshaping enterprise buying.

1. Streamkap: AI workloads broke the case for stale data
For years, the standard objection to real-time data was "we don't need it, a human can read a dashboard once an hour." That argument quietly collapses when the consumer of the data isn't a human.

Streamkap moves data between systems in real time. Their Q1 saw them cross £1m ARR with five customers paying around £100k each, and the founders are explicit that AI agents are the tailwind. Agents and automated workloads consume data continuously and are bottlenecked without it.

The macro picture supports this. Deloitte's State of AI in the Enterprise 2026 report, surveying 3,235 senior leaders, makes the point directly: legacy data and infrastructure architectures cannot power real-time, autonomous AI Deloitte. Their conclusion is that organisations need a "living" AI backbone, real-time, adaptive, and organisation-wide.

Crucially, customers are increasingly choosing to buy this rather than build it. One of Streamkap's larger customers migrated their entire data estate onto the platform in a single week after a competitor outage, a useful signal of how central this infrastructure has become. According to Gartner research cited in recent industry analysis, organisations will abandon 60% of AI projects through 2026 due to the lack of AI-ready data (Cygnet). Companies are realising that fixing the plumbing is the precondition for everything else, and that it's not where they want to spend their engineering headcount.

2. Bourn: distribution beats product
Bourn provides working-capital infrastructure to banks and lenders. In Q1, they signed the second phase of their tier-one bank rollout (NatWest), listed on the Xero marketplace alongside Revolut and Allica Bank, and have now entered the final stages of Microsoft marketplace onboarding.

The strategic move worth flagging is the distribution layer. Rather than only selling to banks, Bourn is embedding into the platforms SMEs already use: accounting software, procurement, marketplaces. McKinsey's analysis of embedded finance projects that, by 2030, embedded channels could account for 20 to 25 percent of retail and SME lending, up from 5 to 10 percent today (McKinsey & Company). That's a structural shift in how financial services reach customers, not a feature war.

The reason customers outsource here is twofold. First, even a tier-one bank can't natively appear inside a customer's accounting software at fintech speed. Second, agentic AI is collapsing the operational cost of running a regulated business. Bourn shipped its first credit-memo agent in Q1, meaning a small team can credibly serve a tier-one bank's customer base without proportional headcount growth. The winning fintechs aren't building the smartest model in market. They're getting into the customer's existing workflow first, and using AI to keep their own cost base flat as they scale.

3. Fulfilment: customers will pay 10x more for an operating system
Fulfilment connects brands with third-party logistics providers. Their recent Haatch investor update contained the most counter-intuitive data point of the three: they introduced subscription tiers starting at £10k and nearly doubled top-line revenue to over £3m, almost entirely from existing customers upgrading.
Same customers, same relationships, materially higher contract values, because their offering moved from "useful tool" to "the platform our workflow runs on." The repositioning is from marketplace to operating system for the 3PL industry.

This pattern is showing up across the market. Bessemer Venture Partners' AI pricing playbook, published in February 2026, observes that in many cases, AI tools replace headcount or augment workflows, reframing the spend narrative from cost reduction to capability expansion (Bessemer Venture Partners). When buyers stop comparing your price to a SaaS tool and start comparing it to an FTE, or to the cost of stitching together a toolchain themselves, the willingness-to-pay shifts by an order of magnitude.

The broader signal: the fastest-growing AI-era B2B companies aren't necessarily those with the most novel models. They're the ones who've successfully repositioned from a feature in someone else's workflow to the system the workflow runs on. Once that switch happens, expansion revenue does most of the work.

What we take from this
Three companies, three different mechanisms: infrastructure that AI demands, distribution that incumbents can't replicate, and operating systems that replace toolchains. None of them are "AI companies" in the headline sense. All of them are seeing sharp growth because AI has changed what their customers are willing to outsource, and what they're willing to pay for it.

The headline numbers tell one story. The reasons behind them tell a more useful one, and it's what we look for when assessing where the next round of growth will come from in the EIS portfolio.