XBE, a software platform for heavy materials, logistics and construction, has secured a growth investment from PSG, although the financial terms of the transaction were not disclosed. The capital will support the company as it expands an AI-driven operating model designed to automate dispatch, logistics and a range of end-to-end operational workflows. Founder and CEO Sean Devine and existing investor Banneker Partners will retain significant ownership positions, with Devine continuing to lead the company.
A Shift From Software Records to Action Systems
The company is moving beyond conventional vertical software that primarily records operational data. It is developing what it calls a "system of action," a platform intended to coordinate and increasingly execute the work that happens around it. That strategy became more visible with the launch of Agent XBE, described as a superworkforce capable of taking operational tasks from beginning to end, and All-22, a ready-mix dispatch simulation and optimization solution.
How Agentic AI Targets Operational Complexity
In construction and heavy materials, an operational AI agent needs access to structured business data, rules, workflows and real-time events. These systems can coordinate orders, dispatch trucks, manage delivery schedules, balance capacity and respond to changing conditions in the field. The appeal is straightforward because thin margins and complex physical logistics still depend heavily on phone calls, spreadsheets, dispatchers and disconnected software systems.
Simulation as a Bridge to Autonomy
XBE's All-22 product is particularly relevant because simulation and optimization provide a controlled environment for testing dispatch decisions before they affect live operations. A poorly optimized dispatch decision can mean an empty truck, a missed delivery, idle workers or wasted fuel, directly connecting software autonomy to margins. If AI agents are eventually making or recommending those decisions, simulation becomes an important bridge between automation and operational trust.
Market Landscape and Competitive Pressures
Deloitte's 2026 construction research found that AI and machine learning tools are already used by 46 percent of surveyed construction businesses, while 50 percent use construction management cloud software. Yet only 16 percent reached an advanced level of overall digital capability, exposing a gap between adopting technology and integrating it effectively. Horizontal vendors such as Microsoft, Oracle and SAP can provide broad AI and workflow capabilities, but vertical specialists can encode industry-specific workflows directly into their products.
Growth Plans and Industry Momentum
With the new investment, XBE plans to recruit aggressively across product, engineering and customer-facing teams, while expanding product development and building its customer community. The company says hundreds of features and enhancements have shipped during 2026, with development shaped directly by customer usage and feedback. It will also prepare to unveil additional products and capabilities at XBE Horizon Live 2026 in San Francisco.
Outcome-Based Software and Governance
Gartner predicts that 40 percent of enterprise applications will incorporate task-specific AI agents by the end of 2026, up from less than 5 percent in 2025. The firm also warns that agentic AI could expose as much as $234 billion of enterprise application software spending to agentic arbitrage through 2030. Adoption will depend on governance as much as automation because only 13 percent of organizations believe they currently have the right controls for AI agents.
XBE's PSG investment positions the company to turn specialized operational data into dependable automation for heavy materials and construction. The larger question is whether vertical software providers can make AI agents reliable enough for decisions that carry physical and financial consequences. XBE is betting they can, and that in this industry the winning software will not merely tell teams what is happening but will increasingly help determine what happens next.