For years, the Silicon Valley narrative around Artificial Intelligence has focused on the "model"—the raw, computational power of Large Language Models (LLMs) capable of writing code, drafting legal documents, and predicting market trends. Yet, as the initial euphoria of the AI revolution settles into the grim reality of corporate implementation, a strange paradox has emerged: the more powerful the AI becomes, the more human labor is required to make it work.
This bottleneck has birthed an entire professional sub-industry of "forward-deployed engineers" (FDEs)—elite technical consultants who are dispatched to corporate offices to manually wire AI systems into legacy infrastructure. It is a costly, slow, and often frustrating patch for a structural problem that is becoming the primary hurdle for the global enterprise.
Enter June, a new startup founded by Salesforce alumni, which emerged from stealth this week with a $20 million pre-seed funding round. The company promises to replace the need for an army of consultants with a platform that maps, cleans, and optimizes enterprise workflows before a single line of AI code is deployed.
The Human-Heavy Reality of AI Deployment
The rise of the FDE is a symptom of a larger, systemic failure. When a Fortune 500 company decides to "go all in" on AI, they quickly discover that their data is not ready. It is fragmented across siloed platforms like Salesforce, Workday, ServiceNow, and Databricks. It is buried under layers of "technical debt"—the accumulated result of years of quick-fix software patches.
Efrat Rapoport, the CEO of June and a former Salesforce executive, characterizes the current industry trend as a "human-intensive pivot."
"AI, paradoxically, increases the demand for professional services," Rapoport explains. "The industry’s answer to AI implementation is, ‘let’s hire more and more and more people.’"
This approach is not only expensive but, according to early adopters, fundamentally flawed. Corporations do not want to be reliant on a revolving door of external consultants who act as a "black box," holding the keys to the company’s most sensitive AI workflows.
A Chronology of the June Origin Story
The genesis of June lies in the previous success of its founding team. Rapoport, along with co-founders Ohad Hen, Barak Goldstein, and Idan Tsitiat, previously built Bonobo AI, a pioneering company that leveraged early language models to transform customer voice-to-text data into actionable business intelligence.
Launched in 2017, Bonobo AI was acquired by Salesforce in 2019. During their tenure at the tech giant, the team spent several years working on the company’s internal AI initiatives. It was during this period that they witnessed firsthand the friction between cutting-edge AI models and the "messy" reality of enterprise software.
"We watched customers struggle to bring AI into their existing platforms," Rapoport says. "We saw them hit the same walls over and over again."
After departing Salesforce, the team decided to tackle the root cause of these failures. Their reputation was such that when they went out to raise capital for their new venture, they didn’t even need a formal pitch deck. The funding round was led by Marc Benioff’s Time Ventures, with high-profile participation from industry titans including Michael Dell, Aaron Levie, and George Kurtz.
The Mechanics of the "Mess": Why AI Projects Fail
To understand why June is necessary, one must look at the "SaaSpocalypse"—the current climate where software firms fear AI might render them obsolete. In reality, no one is "vibe-coding" a complex CRM for a multinational corporation from scratch. Any new AI agent must coexist with the legacy stack.
"Before AI can create value, someone has to deal with legacy systems," Rapoport notes. "You have fragmented data across these platforms. You have complex workflows. You have years of technical debt."
The technical challenge is not in building an AI agent—that is arguably the "easy" part. The difficulty lies in the integration layer. For instance, if a company has 10 duplicate database fields all labeled differently but containing similar information, an AI agent will hallucinate or fail because it cannot discern the "source of truth."
June addresses this by functioning as a diagnostic and architectural layer. The platform scans existing systems to map out business processes and identify the precise bottlenecks where an agent would be most effective. It then provides a step-by-step, automated roadmap for the organization: Remove these duplicates, connect to this specific data source, reconfigure this workflow. Once the environment is "clean," the user clicks a "build" button, and June automates the construction of the agent within the existing infrastructure.
Case Study: The Mortgage Lender’s Dilemma
The practical impact of this technology is best illustrated by the experience of Paul Akinmade, chief strategy officer at CMG, a major U.S. mortgage lender.
Akinmade had ambitious goals to integrate AI into his firm’s operations, specifically aiming to have 100 AI agents active by the time he returned to the Salesforce annual conference. He and his team moved quickly to implement Claude Code, but they hit a wall. Despite consulting with architects and hiring expensive forward-deployed engineers, they remained stagnant for weeks.
"I’ve already done that and I’m getting annoyed by it," Akinmade told Rapoport when considering the June pilot. "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."
By using June, CMG was able to gain a clear, transparent view of where to safely deploy agents. The tool removed the ambiguity that had previously stalled his team’s progress. For executives like Akinmade, June is not just another software layer—it is a replacement for the costly, opaque consultancy models that have dominated the AI implementation phase thus far.
Implications for the AI Industry
The success of June signals a potential turning point in the "AI-as-a-Service" market. If the company’s model proves scalable, it could fundamentally alter the economics of digital transformation.
- The Death of the "Black Box" Consultant: As businesses grow more comfortable with automated diagnostic tools, the reliance on human-heavy FDE firms may wane. This shifts the power dynamic back to the enterprise, allowing internal IT teams to maintain control over their own infrastructure.
- Standardization of Enterprise AI: By providing a structured, step-by-step roadmap for integration, June could help standardize how AI is deployed across different sectors. This would reduce the "wild west" nature of current AI rollouts, where every company is essentially reinventing the wheel.
- The Focus on "Data Hygiene": The rise of platforms like June underscores a crucial truth: AI is only as good as the data it sits upon. The industry is moving away from a fascination with model parameters and toward a sober appreciation for data architecture and cleaning.
- SaaS Integration Stability: Companies like Salesforce and Workday may find that tools like June act as a bridge, making their platforms more "AI-ready" and thus more valuable to clients who are otherwise wary of the integration friction.
Conclusion
As the AI industry matures, the excitement surrounding "agentic" capabilities is being tempered by the reality of corporate bureaucracy and legacy technical debt. The "forward-deployed engineer" was a necessary, albeit unsustainable, response to this friction.
June is betting that the future of enterprise AI lies not in more human labor, but in smarter, automated governance of existing systems. By mapping the messy, fragmented architecture of the modern enterprise, the team is attempting to turn the impossible task of integration into a simple, clickable process. Whether they can succeed in bringing this order to the chaos of the enterprise remains to be seen, but with the backing of tech’s most influential names, the message to the industry is clear: the age of the "manual" AI implementation is coming to an end.
About the Author
