$15M · Atira + 2 more

What they're building

Atira calls itself the commercial brain between a manufacturer's CRM and its ERP, and that framing is precise rather than cute. When a company that builds fire trucks, rail-maintenance machinery or EV chargers is invited to bid, the RFQ that lands in the inbox can run to thousands of pages of specifications, drawings, norms and compliance requirements. Sales engineers, bid managers, legal and commercial staff pass it between them for weeks or months before a price exists. Atira reads that request and runs it through a set of named agents you can see on the product page: requirement extraction, red-flag detection, technical questionnaire, bid/no-bid, norms and standards, commercial feasibility, CPQ, technical documentation, and a system-writeback agent that pushes the output back into SAP, Salesforce and PLM so nobody retypes anything. The design choice that matters is what happens when the documents run out. Every industrial company has one bid manager who has been there since day one and whose retirement keeps the CEO awake at night, and that knowledge is not written down anywhere. Atira's agents ping human experts over Microsoft Teams to resolve the questions no spec sheet can answer, and every generated answer shows its sources and requires a human to confirm each one before it ships. Florian Diegruber's line to Fortune: "If you give Claude to 20 sellers, you still get 20 different answers." The product is not a better assistant for one bid manager. It is a way to make one company's judgment reproducible across the whole org. The traction is unusually concrete for a seed company. Since launching commercially in November 2025 Atira has signed about 15 customers, all of them in full production rather than pilots, with five running more than 100 users each. Chiron Group has 70-plus people on the platform and reports processing inbound quote requests 80% faster. Robel, the railway equipment maker, says Atira saves 95 hours of sales and engineering time per quotation request. ABB E-mobility has deployed it globally, and Rema Tip Top runs it across sales teams in Australia, Brazil and Chile. Deployment takes a day or two and customers can cancel after a month; Diegruber says none has.

Why this matters

Atira estimates industrial sales engineering represents more than $128 billion in annual labour spend worldwide. That is the company's own number and should be read as such, but the shape of the argument holds without it. Manufacturers have spent thirty years automating the factory floor and almost nothing automating the front door. As Diegruber puts it, when your sales cycle takes six months to get from request received to final proposal out, it is nice that you have automated a week of the shop floor, but the big chunk is upstream of it. This work resisted conventional software because the inputs are unstructured, the right answer varies per deal, and the expertise is distributed across people who were never going to fill in a form. The competitive field is real and Atira is not pretending otherwise: Salesforce and ServiceNow are building adjacent quoting automation, the big consultancies are implementing similar tooling, the frontier labs are courting industrial accounts directly, and Roadrunner recently raised $27M from Kleiner Perkins and Founders Fund for the same problem. The counter-argument Atira is running on is that this category is won by whoever can capture tacit knowledge and prove accuracy, not by whoever has the best model access. That is a fight decided by deployment craft, which is why the FDE role is the one to watch.