Magentic, a year old, has closed an $18 million Series A to push AI agents deeper into the procurement and operations work of big manufacturers.
Felicis led the round. Sequoia Capital and The Westly Group, both already on the cap table, joined in.
What Magentic sells is a set of digital workers: multi-agent systems that live inside the software a company already runs, from Microsoft Teams and email to internal enterprise tools. They can carry a job from first question to final invoice, weighing whether to buy or build, picking suppliers, haggling over contracts, tracking orders and settling bills. Scope runs from indirect procurement through to direct spend on raw materials.
The pitch lands on industrial and procurement teams whose workload is climbing while manufacturing demand rises, trade routes move and budgets stay tight. Manufacturers, the company argues, sit on billions of data points, hefty purchasing budgets and a tangle of legacy software and spreadsheets that generic AI tools cannot navigate.
Enterprise deployments come with guardrails: zero-data-retention terms with the big AI providers, room to run in different clouds, and instances isolated by region.
Chief technology officer and co-founder Odhran O’Donoghue said putting advanced AI to work in the physical world takes systems that hold far more context. He described software able to spot a problem, plan a fix, act on it and follow the job through across terabytes of multimodal data.
The new money goes toward building more agents, covering further procurement and supply chain workflows, and research on optimisation problems in industrial operations.
