flowscope is what happens when a Big-Four-style consulting engagement gets compressed into a week by AI agents. Instead of producing a 200-slide deck and a 6-month transformation roadmap, flowscope sends agents into a mid-market or enterprise business to learn how the work actually runs — who touches what, where things stall, and which steps are pure repetition. The output is not a recommendation; it is working automation shipped onto the systems the company already uses. The target buyer is a CFO or COO sitting on top of manual finance, admin, or ops workflows like invoice re-keying, email-based approvals, or weekly spreadsheet rebuilds.
Under the hood, flowscope's agents perform discovery (interviewing operators, reading documents, observing tools), assemble a current-state process map, redesign the flow, and then build and deploy the automations themselves. Because the work runs on the customer's existing stack — ERP, ticketing, email, spreadsheets — there is no rip-and-replace and no long IT integration cycle. The pitch is delivery in days, not months, with consulting-grade documentation as a byproduct rather than the main artifact.
flowscope is in Y Combinator's Spring 2026 (P26) batch, co-founded by Javier Leguina and Samuel Mirpuri. Leguina brings ML engineering experience from a previous startup, and Mirpuri previously worked on AI consulting at McKinsey's QuantumBlack. The team is currently two people based in San Francisco, and the product is sold as bespoke engagements rather than self-serve SaaS.