Skan AI raised $63 million in Series C funding, co-led by Cathay Innovation and Dell Technologies Capital, to scale a platform that watches employee screens in real time and builds a working model of how enterprises actually operate. Citi Ventures, Bloomberg Beta, State Farm Ventures, and Wipro Ventures also participated, bringing total funding to roughly $120 million. The company is simultaneously launching general availability of two products: Skan AI Blueprint and Skan AI Agents, which sit on top of its existing intelligence layer to form a full pipeline from process discovery to automation.
The funding thesis rests on a documented failure rate. Gartner data shows only 8% of enterprises have AI agents in production, and 95% of early implementations will require a complete redesign. Co-founder and CEO Avinash Misra argues the cause is not the models but the inputs. Process documentation and system logs only capture committed states of work. The 80% that happens between those states, the exceptions, the rework, the cross-application handoffs, never appears in the data that most companies are feeding their agents. Skan deploys observation technology on employee desktops to capture that gap directly, then abstracts it into a live process model AI can reason over. The hard part, Misra is careful to note, is not screen observation but intent extraction: teaching a model to distinguish a new case from rework on an existing one, statefully, across 1,500 workers simultaneously.
The full article is worth reading for two reasons. First, Misra's technical argument against process mining vendors like Celonis is precise and cuts at an assumption the entire industry has accepted. Second, the surveillance question is not dismissed but architecturally addressed: Skan aggregates statistical patterns across hundreds of workers performing the same role, not individual behavior logs, a design choice that originated from a direct challenge by an early enterprise customer. Both threads matter for anyone evaluating where enterprise AI actually breaks and what fixing it requires.
[READ ORIGINAL →]