Forgepoint raises $9.8M to automate US manufacturing sales
New York-based Forgepoint has closed a $9.8 million seed round to scale Frank, an AI agent designed to automate the quote-to-order workflow for American manufacturers. The round was led by Primary Venture Partners and a16z speedrun, with angel participation from figures at OpenAI, Brex, Vercel, and SoFi. The raise arrives at a moment when US industrial policy is pushing manufacturers to expand output, yet the back-office processes that determine whether a factory wins a contract have seen little of the automation lavished on the shop floor itself.
The core thesis is straightforward: a manufacturer's revenue capacity is constrained not by its machines but by how quickly its team can price and return a customer's request for quotation. A 2025 Gardner Intelligence study cited by Forgepoint found that nearly half of manufacturers expect a vendor response within 24 hours, a threshold that is routinely missed when a single estimator must parse email attachments, build a cost model, chase missing specifications, navigate an ERP system, and manage follow-up across multiple open jobs simultaneously.
From the shop floor to the inbox
Frank is configured to slot into each manufacturer's existing systems and operating procedures rather than replacing them. It reads inbound RFQ emails, flags incomplete specifications, contacts buyers for clarification, builds a draft quote using the shop's own cost model, routes it for human sign-off, and manages follow-up correspondence after approval. The agent is domain-agnostic within manufacturing: Forgepoint cites deployments across metal service centres, foundries, injection moulding, fabrication, and precision machining, with configuration varying by shop type.
Forgepoint says customers have increased monthly quote throughput by an average of 30% and cut time-to-quote by more than 55% since deployment. The company reports reaching seven-figure ARR within six months of Frank's launch. These figures come from the company's own release and have not been independently verified.
CEO Jack Strauss brings an industrials background spanning Northrop Grumman, Bain's Advanced Manufacturing practice, and aerospace integrator Xwing. Co-founder and CTO Alex Pareto previously led engineering for Brex's spend-management platform. The pairing is notable: one founder with deep manufacturing-floor credibility, the other with enterprise software architecture experience, a combination that reflects the genuine difficulty of building AI agents that must learn idiosyncratic, shop-specific pricing logic rather than operate on clean, standardised data.
The broader convergence play
The strategic context extends well beyond a single seed round. US industrial policy, from the CHIPS Act to the broader reshoring agenda, has directed significant capital toward expanding domestic manufacturing capacity. That push creates a structural demand problem: factories can be built or retooled, but the administrative workforce needed to handle the resulting surge in customer enquiries does not scale at the same pace. Agentic AI, applied to quoting, scheduling, and procurement workflows, is positioned as the mechanism that allows manufacturers to grow revenue without proportional headcount growth. It is a labour-substitution thesis wrapped in an industrial-competitiveness narrative, and it is attracting capital accordingly.
The investor syndicate here is itself a convergence signal. a16z speedrun targets early-stage AI-native companies; Primary Venture Partners has a record in enterprise software; the angel cohort from OpenAI and fintech-adjacent firms such as Brex and SoFi suggests a broader bet that agentic workflow tools will become horizontal infrastructure across multiple industries. Manufacturing quoting is the beachhead, but Forgepoint has signalled an intention to extend Frank's model across other back-office functions, from procurement and planning through to finance operations.
For cross-sector investors, the question is whether the RFQ workflow is a sufficiently large and defensible wedge, or a feature that larger ERP and CPQ (configure-price-quote) vendors will eventually absorb. Established platforms such as SAP, Oracle, and Salesforce all operate in adjacent territory. Forgepoint's answer, for now, is that enterprise software was built for standardised processes and cannot easily accommodate the craft pricing logic that lives inside individual manufacturing shops. Whether that moat holds as foundation models improve will be the defining test of the next funding round.