Every email an employee sends is an unused brand and marketing channel. WiseStamp turns signatures into a centrally managed asset across the whole company.

Reco is an AI Agent Security Platform. Built on the Reco Graph, the company helps organisations discover agents and AI-enabled tools, prioritise risk and remediate with precision across agents, identities, applications, permissions, data, and workflows. With more than 280 app integrations and 1,000 detection controls, Reco helps enterprises govern agents wherever they operate.

As AI agents become embedded across enterprise applications and workflows, organisations need visibility into which agents are active, what identities and permissions they use, and what data and systems they can access. Reco addresses this emerging security challenge with a graph-based platform that helps enterprises reduce risky access while supporting legitimate AI adoption.

Fonoa is the AI-powered Tax Operating System for autonomous tax. Fonoa’s modules cover the full indirect tax lifecycle: tax ID validation, real-time tax determination, e-invoicing, and returns. Leading global companies such as Canva, Uber, Netflix, Nebius, and Booking.com all work with Fonoa to navigate complex, real-time tax requirements as they scale across markets, seeing up to 90% faster tax calculation in some cases.

Fonoa is leading the shift in tax from a series of disconnected point solutions to a unified operating system. The category has historically spanned tens of billions globally, including services now being reshaped by automation, and Fonoa is enabling businesses to manage compliance with a level of speed and intelligence that wasn’t previously possible.

Procurement software has been slow, ugly, and unloved for two decades. Pivot rebuilt it as a consumer-grade product that finance and employees actually want to use, with AI compressing the approval and sourcing loop. Enterprise spend is one of the largest untouched surfaces for applied AI.

Turning a phone into a payment terminal removes the last piece of hardware from commerce. Mobeewave built the software to do it securely, and the platform owners noticed. Infrastructure this fundamental rarely stays independent for long.

The intelligence that lets a robot move safely through a factory is the hardest part to build. BlueBotics licensed that navigation layer to the industry. Foundational IP that others depend on is a quietly powerful place to sit.

IT has always measured servers, never the human on the other end of the laptop. Nexthink instrumented the employee experience across millions of endpoints. A category created and then owned – the outcome we build toward.

A brain scan holds more than any radiologist can measure by eye. Icometrix quantifies it, turning images into clinical-grade markers for MS, dementia, and beyond. Regulated, data-rich AI in healthcare is slow to build and hard to displace.

Hotels sell the same room through dozens of channels at prices they can barely see. Fornova gives them a single view of demand, rate, and distribution. Category-defining data in a fragmented vertical is exactly where focus pays off.

The smallest businesses are the least served by software and the most numerous. vcita bundles scheduling, payments, and marketing into one system a solo operator can run. AI turns that suite into an operator that works while the owner sleeps.

AI’s appetite for compute is bottlenecked by how fast chips can talk to each other. Kandou’s interconnect moves more data with less power at the physical layer. Deep-tech IP that sits on the critical path of the AI build-out is rare and defensible.

Enterprise data is scattered across systems that were never meant to talk. K2view organises it around the business entity – the customer, the order – and serves it in real time. The plumbing beneath every AI initiative is a market that only grows.

Getting a thousand people to think together usually collapses into noise. Howspace uses AI to synthesise, cluster, and steer large-group dialogue in real time. Software that makes human coordination legible is quietly one of the biggest enterprise categories.

A broken-down bus or truck is a data event that could have been predicted. Stratio reads the vehicle in real time and calls the failure before it happens. Sensor data at fleet scale is a compounding advantage no incumbent can copy overnight.

Every parcel, shelf, and asset is a data point waiting to be read by the camera already in someone’s hand. Scandit turns any device into an enterprise-grade scanner. Computer vision embedded in the physical operations of retail and logistics is applied AI at scale.

Matching shift work to willing workers is a logistics problem dressed as an HR one. Zenjob automates the entire loop – sourcing, scheduling, payroll – into a few taps. Marketplaces with a real software spine defend themselves.

Shipping a painting or a sculpture is a bespoke, high-stakes operation run on spreadsheets and phone calls. Convelio built the pricing and operating layer that makes it instant and insurable. Vertical software that owns a hard physical workflow is a durable position.

When a number is wrong, the hardest question is where it came from. Manta mapped data lineage automatically across the tangle of enterprise systems. Governance infrastructure becomes non-negotiable the moment AI touches production data.

AI runs on data, and data has to live somewhere affordable. Wasabi undercut the hyperscalers on object storage without the egress penalties. Infrastructure priced to win volume is a structural advantage in a market that only expands.

Security teams drown in tools that each see a slice of the risk. Veriti unified posture and safe remediation across the stack. Consolidation plays in security, done with real data, compound into platforms.

Contracts are where every enterprise obligation lives, and reading them is still done line by line. Luminance built legal-grade models that surface, negotiate, and redline at a lawyer’s standard. Deep domain data plus a specialised model is the pattern we keep returning to.

Pricing is the highest-leverage decision a company makes and the least instrumented. Buynomics simulates millions of virtual customers to answer what a change actually does before it ships. It replaces gut feel with a repeatable model – the kind of workflow AI was built to own.

Physical products are still designed in slow simulation loops that take days per iteration. Neural Concept learns the physics and returns answers in seconds, collapsing the engineering cycle for cars, aircraft, and hardware. Applied AI at the core of how things get built.

Consumer brands fly blind between the shelf and the checkout. Datavations assembles the missing data layer – pricing, assortment, and share – and makes it decision-ready. Proprietary data feeds that improve with scale are the moats we look for.