Partnering with Magentic: AI-Driven Savings for the World’s Supply Chains
Sequoia by Julien Bek and Zefi Hennessy Holland
AI agents are moving into procurement, not just chat. Magentic says its bots can find savings in supplier chaos that humans can’t.
Based on reporting by Sequoia, by Julien Bek and Zefi Hennessy Holland — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Three years after the ChatGPT moment, enterprise AI is starting to shift from answering prompts to doing the work. Sequoia says the business model is shifting too: instead of selling software seats and hoping the ROI shows up later, companies are being pushed to sell outcomes directly.
That matters most in procurement, where most company costs pass through and the teams are tiny compared with the supplier sprawl they’re meant to police. Sequoia points to large enterprises working with tens of thousands of suppliers while their procurement function is only 1/100th that size. McKinsey’s estimate, as Sequoia cites it, is blunt: unfulfilled supplier obligations can leak roughly 2% of spend, which for some customers means tens of millions lost every year.
Magentic is trying to close that gap with what it calls “Mages.” These AI agents plug into existing ERPs, work alongside procurement teams and comb through thousands of documents and millions of deliveries. The pitch is not just detection. The system is meant to spot late payments, quality issues and other problems, trace the chain of events behind them, and then suggest next steps for a human to approve, like drafting emails, updating invoices or alerting stakeholders.
Sequoia is framing this as a different kind of automation. The point is not to replace procurement staff doing ordinary work. It is to handle the sort of scale and pattern-matching that people simply cannot do by hand, while keeping humans in control. That’s also the reason the fund says it wants to back vertical agents: they can sell into software budgets, sure, but they can also take a cut of the savings they create.
The founders fit the story Sequoia wants to tell. Robin Van Aeken and Odhran O’Donoghue met while winning hackathons at Oxford, with Odhran coming from OpenAI and Robin bringing McKinsey procurement experience. Sequoia says it met them through Arc, its pre-seed and seed immersion program, and is now leading their seed round.
My take — AI-written commentary, not fact-checked reporting
This is the part of AI that deserves the attention: boring, expensive back-office work where a model can point to cash, not vibes. Procurement is a perfect target because the waste is real and the org chart is tiny, which is exactly why the usual enterprise software sales pitch feels tired here. If AI ever earns its keep, it’ll be by catching money leaking through the floor, not by writing nicer email drafts.
Read more about this at: Sequoia