Conxai Raises €5M to Automate Construction’s Data Problem

Construction runs on data. It always has. Drawings, submittals, RFIs, camera feeds, sensor readings, daily reports – a mid-size project generates more information than most teams can realistically process. The problem is not a shortage of data. It is that most of it sits untouched, split across tools that do not talk to each other, formats that require manual handling, and workflows that depend on someone with fifteen years of site experience to interpret correctly.

That is the gap Munich-based ConXai is going after – and investors just backed them to the tune of €5M to do it. The round, led by BayBG Venture Capital and Capricorn Partners, brings ConXai’s total raise to around €7.7M since its 2022 pre-seed with Earlybird, Pi Labs, noa, and Argonautic Ventures. Zacua Ventures, one of the more construction-specific VC firms out there, also participated.

The timing is deliberate. A ConTech Investor Survey published by Zacua Ventures in February 2026 – drawing on responses from 140 global investors – found that 84% intend to maintain or increase capital deployment in the sector this year. Agentic AI is the specific category they are circling. ConXai’s raise lands squarely in that window.

What ConXai Actually Does

Purpose-built, not adapted

The distinction ConXai pushes hardest is that its platform was trained on construction workflows, data structures, and industry processes from day one – not a general-purpose model retrofitted for AEC. That might sound like marketing framing, but it matters in practice. Off-the-shelf tools struggle to interpret a fragmented tender submission or cross-reference a camera feed against a project schedule. ConXai’s system is built to handle exactly those tasks.

The platform pulls from a wide range of unstructured sources – photographs, video footage, sensor readings, CAD files, and project documents – and uses them to automate reporting, document processing, and project control workflows. A module called SiteLens handles real-time site visibility, tracking labour, equipment utilisation, and conditions without someone manually compiling that picture at the end of each day. On the commercial side, it can analyse tender submissions, extract key technical and commercial information to speed up bid levelling, and flag risks before they become problems.

The no-code setup means project teams can configure use cases without needing engineering resource – which, for contractors who are not running software teams, is genuinely the difference between deploying a tool and watching it gather dust. It is the same friction point we have covered in how AI is starting to cut through construction’s document review bottleneck – the technology is only useful if teams can actually run it.

The Architecture Behind It

Agents that act, not just advise

The technical term ConXai uses internally is Neuro-Agentic Reasoning Architecture – which is a dense way of saying the system combines what it knows from training with real-time multi-modal input, then produces automation that is auditable and explainable rather than a black box. The outputs adapt to context rather than following fixed rules. That matters in construction, where no two projects look the same and conditions change week to week.

The practical implication is that the agents are designed to execute workflows, not just surface information. That is a meaningful distinction from most AI tools being marketed into construction right now, which still require a human to take the data and act on it. As adoption patterns show, AI adoption across the trades is moving faster than most expect – but the tools that stick are the ones that reduce manual load rather than add another dashboard to monitor.

Faisal Butt, managing partner at Pi Labs and an early backer of ConXai, described the core value plainly: construction generates vast amounts of data that has historically been unusable, and the platform turns fragmented information into fast, high-quality decisions. Steven Lambert at Capricorn Partners pointed to efficiency gains, improved risk management, and better transparency as the headline outcomes. Neither is a speculative pitch – those are the exact pain points contractors face on live jobs right now.

Where the Money Goes

North America, Europe, and Asia on the roadmap

ConXai’s stated plan for the fresh capital covers three areas: continued platform development to expand into additional workflows beyond the current feature set, and a push into North America, Europe, and Asia. The company is currently headquartered in Munich with its primary traction in European construction markets. Breaking into North America is a different sales challenge – longer sales cycles, different procurement norms, and a more fragmented contractor landscape than in Germany. That expansion will be a real test of whether the product travels or whether it needs significant reconfiguration for different markets.

McKinsey’s estimate that construction’s inefficiency amounts to losses in the trillions globally gives investors a rough sense of the addressable problem. The more specific data point – that around 30% of project data is typically lost once a construction project completes, even when it could inform future work – speaks directly to what ConXai is trying to fix. Not just within a project, but across the knowledge base a firm builds up over years of delivery.

The Broader Signal

Agentic AI is the category investors are betting on

What ConXai’s raise confirms is less about a single company and more about where conviction is sitting in the ConTech market right now. Agentic AI – meaning systems that complete tasks rather than assist with them – is the specific framing investors are responding to. The general “AI for construction” pitch has been around long enough that it needs specificity to cut through. Vertical platforms trained on construction data, with clear workflow automation and measurable output reductions, are what is getting funded.

Whether ConXai can execute at the pace the raise implies is a separate question. The product exists, it has real backers with genuine sector knowledge, and the problem it is targeting is not going away. Construction firms are sitting on data they cannot use, running workflows that still rely heavily on manual judgment, and operating on margins thin enough that any meaningful efficiency gain compounds quickly. That combination – real problem, existing product, specialist investors – is a reasonable foundation. The North America expansion will tell you a lot more about whether the thesis holds.

Related Articles
How AI Is Automating Construction Document Review
78% of Trade Contractors Use AI – AI Adoption in Construction Is Way Ahead of What You Think