Across AI, a new startup pioneering Agentic AI for enterprise workflows, announced it has raised an oversubscribed seed round to bring to market the industry’s first enterprise sales AI supported by next generation agentic memory. Across AI’s groundbreaking technology, based on years of academic research, reasons over internal and external enterprise data to provide accurate, verifiable insights and actionable recommendations for entire business workflows, starting with enterprise sales. The funding was co-led by Village Global, a venture firm backed by some of the world’s most successful entrepreneurs, including Mark Zuckerberg, Eric Schmidt, and chaired by Reid Hoffman, and also co-led by early Dropbox investor Bobby Yazdani at Cota Capital.
Across AI: Organizing enterprise knowledge with an agentic memory system to proactively augment B2B sales teams.
The Next Generation of Agentic Memory
Despite advances in AI, supporting complex enterprise workflows remains a challenge. Current approaches struggle to understand the nuanced context and relationships within enterprise data. Techniques like Retrieval-Augmented Generation (RAG) are useful for answering factual questions, but they treat all data equally and lack the ability to prioritize, reason over, or adapt to new and changing information. These limitations prevent existing solutions from reliably performing the high-level reasoning needed to transform enterprise workflows.
Across AI’s technology overcomes these challenges with a dynamic memory system purpose-built to support complex workflows. Unlike first-generation knowledge retrieval systems, Across AI’s memory continuously adapts, retaining only relevant information, discarding outdated data, and surfacing key insights based on contextual importance. This enables Across AI’s agentic AI apps to reason deeply over unstructured enterprise data within each step of a workflow by using a proprietary combination of large language models (LLMs) and workflow-specific algorithms.
Existing sales AI solutions may handle simpler tasks but often fall short due to their “AI-wrapped” nature—superficial layers over traditional systems that lack deep integration and contextual understanding. Enterprise sales teams have to execute complex and interconnected sales workflows. Across AI addresses this challenge with Agentic apps that support sales teams across each sales stage and are built for multi-step workflows. They synthesize data into higher-level reasonings, which are used to perform accurate actions such as generating a document.
“Having carried a quota for 18 years, I’ve lived the daily challenges salespeople face,” said Steven Mih, CEO and Co-founder of Across AI. “Throughout my career, deep product and account knowledge gave me an edge in prospect conversations but took an enormous amount of time. As a sales manager, I depended on each team member to build a pipeline and execute processes at the highest level. We started Across AI to build the solution I always wished I had—that understands the context of my deals and has the collective knowledge of my sales team to assist me with expert-level strategy and tactics.”
Meet the Founding Team
- Steven Mih, CEO and co-founder, is a repeat entrepreneur with decades of enterprise sales experience. He was previously the co-founder and CEO of Ahana Cloud, which was acquired by IBM in 2023. Steven has led successful startups as the former CEO of Aviatrix and has led sales teams at Couchbase, Transitive, and Cadence.
- Dr. Niloufar Salehi, CPO and co-founder, is an AI industry-known researcher in human-centered AI. She has led groundbreaking research on AI system design as a professor at UC Berkeley, shaping how AI can be integrated into complex workflows to boost reliability and augment human capabilities.
- Dr. Afshin Nikzad, CTO and co-founder, is a specialist in advanced algorithms and machine learning models. He is a dual PhD recipient from Stanford University, a competitive programming gold medalist, and is on leave from his professorship at the University of Southern California (USC). His expertise enables Across AI’s agents to process and understand nuanced contexts and provide actionable insights that adapt to new information in real-time.
Based on UC Berkeley research and NSF grants, the team is uniquely positioned to tackle the critical challenges in today’s sales technology landscape.
Read More: SalesTechStar Interview with Christina Brady, CEO of Luster AI