For AI to work reliably in big enterprises, a new layer of professional service—forward-deployed engineers (FDEs)—has emerged to fight with legacy systems, fragmented data, and technical debt. A Marc Benioff-backed startup called June argues that hiring more FDEs is not the only path forward. Founded by four former Salesforce executives who previously built Bonobo AI, June exited stealth on Monday with $20 million in pre-seed funding from Time Ventures, Michael Dell, Aaron Levie, and George Kurtz. The company declined to share its valuation.

June‘s premise is that any AI model must integrate with platforms such as Salesforce, ServiceNow, DataBricks, and Workday before it can produce value. Building an agent template is easy; getting it to work with duplicate database fields, inconsistent team workflows, and years of accumulated debt is the real bottleneck. June scans a company’s existing systems to map business processes, surface bottlenecks, and generate an automated, step-by-step implementation roadmap for agent-powered alternatives. The platform notifies teams via existing communications channels and guides them through tasks such as removing duplicates or connecting new data sources. Once steps are defined, users click a build button and June constructs the agent within the organization.

CMG, a major U.S. mortgage lender, tested June after its chief strategy officer Paul Akinmade found that moving to Claude Code was straightforward but integrating with Salesforce stalled his team for weeks. Traditional consulting and FDE conversations yielded no progress. June provided a clear view of where to deploy agents and allowed safe implementation before the official kickoff call between the two companies. Akinmade told Rapoport he wanted a tool that did not require FDEs or a black box—June delivered.

Rapoport sees June as complementing FDEs and consultants, but customer interest appears driven by the desire to bypass them entirely. The company’s go-to-market strategy and pricing were not detailed in the announcement.