YC’s RamAIn Takes Aim at the ‘Legacy Gap’ With High-Speed Computer-Use Agents
Table of Contents
Bridging the Gap Between Modern AI and Legacy Software
The central tension in enterprise digitisation has always been the ‘legacy gap’—the friction between cutting-edge AI capabilities and the clunky, decades-old desktop applications and web portals that still power global commerce. While LLMs can draft emails and write code, they generally struggle to interact with the specific, non-API-driven interfaces of legacy enterprise software. RamAIn, a new addition to the Y Combinator Winter 2026 cohort, is attempting to solve this by building what they describe as the world’s fastest ‘computer-use agents.’
Unlike traditional Robotic Process Automation (RPA), which often relies on rigid, brittle scripts that break the moment a UI element shifts by a few pixels, RamAIn is developing agents that operate with human-like intuition. The goal is to enable an AI to navigate a screen, interpret visual cues, and execute complex tasks across desktop apps and portals using natural language commands, effectively treating the entire operating system as its interface.
Founding Pedigree: From IIT Delhi to CMU
The company is led by a duo combining deep operational insight with high-level machine learning research. CEO Shourya Vir Jain brings a perspective shaped by his tenure at McKinsey, where he observed the sheer volume of manual labor required to keep legacy enterprise systems functioning. This operational pain point, coupled with a background in scaling a previous enterprise AI venture to six-figure ARR, provides the business logic for RamAIn’s trajectory.
On the technical front, CTO Vansh Ramani provides the architectural muscle. A researcher previously affiliated with Carnegie Mellon University (CMU) focusing on scalable machine learning and representation learning, Ramani has a track record of optimizing how machines process data. His work includes publications at ICLR and ACS, and notably the development of “Panaroma,” a vector search algorithm that was eventually merged into Meta’s FAISS (Facebook AI Similarity Search), one of the industry standards for efficient similarity search in large datasets.
This combination of McKinsey-style workflow analysis and CMU-level algorithmic optimization is a strategic bet on the ‘reasoning and planning’ layer of AI. For RamAIn to succeed, their agents cannot simply predict the next token; they must plan a sequence of actions across a visual interface and correct themselves in real-time when a legacy system lags or throws an unexpected error.
The Race for the ‘Action’ Layer
RamAIn enters a crowded but critical race. The industry is currently shifting from ‘Chatbots’ (which provide information) to ‘Agents’ (which execute actions). With Anthropic recently debuting its ‘computer use’ capability for Claude, the battleground has shifted to who can make these interactions reliable and fast enough for an enterprise environment.
Enterprise clients are notoriously risk-averse. An AI agent that makes a mistake in a CRM is a nuisance; an agent that makes a mistake in a legacy financial ledger or a supply chain portal can be catastrophic. RamAIn’s focus on being “10x faster and more reliable” suggests they are targeting the latency and accuracy issues that have plagued early computer-use prototypes.
Expanding the Team
As part of its current growth phase, RamAIn is moving beyond its founding core to build out its go-to-market (GTM) infrastructure. The company is currently seeking its first non-founding business hire—a technical GTM builder. This role is specifically designed to bridge the gap between the product’s technical capabilities and the revenue pipeline, focusing on building AI-driven outbound systems and personalization pipelines.
This hiring move signals that the company has moved past the initial R&D phase and is now focused on validating its product-market fit within the enterprise sector. By automating the very top-of-funnel activities they are hiring for, RamAIn is essentially using its own philosophy of AI-native automation to scale its business operations.
