Large companies are finding that getting AI tools into daily operations is not just a model problem. It is an implementation problem. June, a new startup from former Salesforce executive Efrat Rapoport and three cofounders, is betting that AI itself can help solve that bottleneck.
The company emerged from stealth Monday morning with $20 million in pre-seed funding led by Marc Benioff’s Time Ventures. It says its platform can examine a company’s existing systems, identify workflow problems and help build agent-powered processes that can actually run inside an enterprise environment.
Enterprise AI is running into the old software stack
The rise of AI has created a new kind of services layer around deployment. Forward-deployed engineers, or FDEs, are specialists who go directly into companies to make AI systems work in real settings. Their growth reflects a basic reality: many businesses cannot simply plug a model into their operations and expect reliable results.
Rapoport frames the problem as a paradox. “AI, paradoxically, increases the demand for professional services,” she says. In her view, the current industry answer has often been to add more people around the implementation challenge.
That challenge starts with the systems companies already depend on. Any AI agent used in a large business may need to work with platforms such as Salesforce, ServiceNow, DataBricks, Workday or other data-management tools. Those platforms hold critical business context, but they also come with fragmented data, complicated workflows and years of technical debt.
Rapoport’s argument is that the easy part is creating an agent template. The difficult part is making that agent useful inside a company where different teams may use overlapping fields, duplicate records or inconsistent processes. As she puts it, an agent has to know what to do when there are “10 duplicate [database] fields that say the same thing, and different teams are using them.”
What June is trying to automate
June’s product is designed to map the environment before deployment. According to the company, the platform scans existing systems to understand business processes, spot bottlenecks and design improved workflows powered by agents. It can also notify teams through the company’s communications channels.
The goal is not only to recommend where AI should go, but to produce a practical sequence for getting there. Rapoport says June gives customers “the full roadmap automatically of what needs to happen step by step” so an agent can be implemented successfully in a complex enterprise environment.
That roadmap can include concrete tasks such as removing duplicates or connecting to a data source. The company’s pitch is that customers can then click “build” on each task and have June begin creating the needed pieces inside the organization.
In plain terms, June is positioning itself between the promise of AI agents and the operational reality that blocks many deployments. It is not arguing that companies lack interest in AI. It is arguing that the path from interest to working deployment is filled with system cleanup, integration work and process redesign.
The Salesforce link behind the founding team
Rapoport started June with Ohad Hen, Barak Goldstein and Idan Tsitiat. The four founders previously built Bonobo AI, a pre-transformer language model company that launched a voice-to-text service in 2017.
Salesforce acquired Bonobo AI two years later. The team then spent several years working on the company’s AI initiatives. That experience shaped June’s focus, according to the source article, because the founders watched customers struggle to bring AI into existing platforms.
The company attracted notable backers for its pre-seed round. In addition to Time Ventures, investors include Michael Dell, Aaron Levie and George Kurtz. June declined to share its valuation.
Rapoport says investor interest came quickly enough that “we didn’t even have a deck for this raise.” The funding gives June a chance to test whether enterprises will prefer software that guides and builds AI deployments over heavier reliance on outside implementation teams.
A customer case shows the pain point
One early example in the source article comes from CMG, described as a major U.S. mortgage lender. Paul Akinmade, the company’s chief strategy officer, had moved CMG’s software engineering over to Claude Code quickly. But integrating that work with Salesforce became a roadblock.
The timing mattered because Akinmade had said at Salesforce’s annual conference the year before that he would come back with 100 agents running. As the integration problems continued, that goal looked harder to reach.
Akinmade says his team spent weeks trying to make progress, including meeting with architects, speaking with forward-deployed engineers and consulting others. June, he says, helped the team see where to deploy agents and do it safely, even before the official kickoff call between the two companies.
His concern was not only whether the technology worked. It was whether using it would require another layer of specialists. When he was considering a pilot, he told Rapoport: “If your product requires FDEs, I don’t want your product. I’ve already I’ve already done that and I’m getting annoyed by it. I don’t want a black box. I don’t want something only certain people can figure out. I want an easy-to-use tool.”
The bigger question for AI deployment
June arrives as software companies debate how AI will reshape the market. The source article notes that fears around a so-called SaaSpocalypse have raised the possibility that AI could replace software firms. But it also points out that no one is casually building a CRM for a Fortune 500 company with AI alone.
That distinction matters. AI may generate code, create agents and speed up engineering work, but enterprise adoption still depends on the systems, data and workflows already inside a company. A useful agent has to operate in that environment, not in a clean demo.
Rapoport sees June as a tool that can complement FDEs and consultants. Some customers, however, may value it because it could reduce the need for them. That tension is central to June’s pitch: enterprises want the benefits of AI, but they also want deployment to be understandable, repeatable and not limited to a small group of specialists.
If June can deliver on that claim, its role would be less about inventing another AI use case and more about making existing AI ambitions easier to execute. For many large companies, that may be the part of the AI stack that determines whether agents remain experiments or become part of everyday operations.