The Anti-Consultant Play: June Bets on Automation to Deploy Enterprise AI

The Paradox of AI Implementation

Large enterprises face a stubborn contradiction: tools designed to automate work now require entire teams of specialists to get them running. Forward-deployed engineers, or FDEs, have become a fixture in corporate AI adoption, parachuting in to wire up models to Salesforce instances, untangle duplicate database fields, and navigate the technical debt accumulated over decades. The arrival of generative AI has not reduced this dependency. It has deepened it.

June, a startup that exited stealth mode this week, is building software to make those specialists obsolete. The company secured $20 million in pre-seed funding led by Time Ventures, the investment vehicle of Salesforce founder Marc Benioff. Michael Dell, Box CEO Aaron Levie, and CrowdStrike CEO George Kurtz also participated. June declined to disclose its valuation.

The founding team - Efrat Rapoport, Ohad Hen, Barak Goldstein, and Idan Tsitiat - previously built Bonobo AI, a pre-transformer language model startup that Salesforce acquired in 2019. During their years inside Salesforce, they watched customers struggle to integrate AI into production environments. The challenge was not the models themselves. It was the infrastructure beneath them: fragmented data, incompatible workflows, and platforms that had evolved through years of mergers and patched-together integrations.

Rapoport, who now leads June as CEO, noted that the pitch to investors required no formal deck. The team's track record and the problem's visibility were enough.

Mapping the Mess Before Building Agents

The core obstacle in enterprise AI deployment is not generative capability. It is operational legibility. Companies run on systems that were never designed to interoperate - Salesforce for customer data, ServiceNow for IT operations, Databricks for analytics, Workday for HR. Each platform holds fragments of the same information, often duplicated or contradictory. An AI agent tasked with automating a sales workflow might encounter ten different fields labeled "customer tier," each used by different teams with different definitions.

June's software begins by scanning a company's existing technology stack. It identifies business processes, flags bottlenecks, and surfaces redundancies. Then it generates a step-by-step roadmap for deploying AI agents, complete with instructions to consolidate duplicate fields, connect data sources, and reconfigure workflows. Teams receive notifications through existing communication channels. Once the roadmap is in place, users can click through tasks, and June builds the integrations automatically.

This approach inverts the usual sequence. Instead of hiring consultants to audit systems and then hiring engineers to implement fixes, the platform performs both functions. The aim is to compress months of discovery and integration work into a software-driven process that requires minimal human intervention.

A Mortgage Lender's Test Case

Paul Akinmade, chief strategy officer at CMG, a major U.S. mortgage lender, moved his engineering team onto Anthropic's Claude for code generation without difficulty. The friction appeared when he tried to integrate the AI with Salesforce. Akinmade had publicly committed to deploying 100 agents by a deadline tied to Salesforce's annual conference. Weeks of meetings with architects and FDEs produced no progress.

June provided a different path. The platform mapped CMG's existing systems, identified where agents could operate safely, and generated deployment instructions. Akinmade's team cleared the bottleneck. In conversations with Rapoport, he set a clear condition: if the product required forward-deployed engineers, he was not interested. He wanted a tool his team could operate independently, without relying on external specialists or opaque processes.

That requirement reflects a broader tension in enterprise AI. FDEs are expensive, their availability is constrained, and their work often leaves companies dependent on outside expertise. June positions itself as a complement to these specialists, but its appeal to customers like CMG lies in its potential to eliminate the need for them altogether.

The Professional Services Trap

The rise of AI has paradoxically expanded demand for human labor in implementation. Software vendors that once sold licenses and left customers to configure systems now send teams of engineers to ensure deployments succeed. This model generates revenue, but it also creates friction. Projects stretch across quarters. Costs accumulate. Internal teams lose autonomy.

Rapoport described this dynamic as unsustainable. The industry's default response to AI complexity has been to scale up professional services - hiring more consultants, more FDEs, more integration specialists. June's thesis is that the problem is structural, not one of insufficient labor. The real issue is that enterprises lack visibility into their own systems. They do not know where data lives, how workflows connect, or which processes are candidates for automation.

By automating the discovery and integration phases, June aims to collapse the dependency on external expertise. The platform does not replace the need for technical judgment, but it shifts that judgment from external consultants to internal teams equipped with better tools.

The Competitive Landscape

June enters a crowded field. Startups like Gleen, Kore.ai, and Moveworks are building agent orchestration platforms. Salesforce itself offers Agentforce, a suite of pre-built agents designed to integrate with its ecosystem. ServiceNow and Microsoft have similar offerings. The question is whether enterprises will adopt vendor-specific agent frameworks or seek tools that work across platforms.

June's bet is on the latter. Its value proposition depends on being platform-agnostic - able to scan and integrate with any combination of enterprise software. That positions it as infrastructure rather than a vertical solution. The company is not building agents for specific workflows. It is building the layer that makes those agents functional in real-world environments.

The funding round signals investor confidence in this approach. Time Ventures, Dell Technologies Capital, and executives from Box and CrowdStrike bring both capital and distribution potential. Benioff's involvement is particularly notable given his role at Salesforce, a company that sells both AI tools and the professional services to deploy them. His backing of June suggests a recognition that the current model is unsustainable, even for incumbents.

What This Means for Enterprise AI Adoption

If June's approach works, it could accelerate AI adoption by removing a significant bottleneck. Enterprises that have hesitated to deploy agents due to integration complexity might find the process more tractable. Internal teams could take ownership of AI deployments without waiting for external specialists. The cost structure of AI implementation could shift from labor-intensive consulting to software subscriptions.

The risk is that enterprises are not ready to operate without human guidance. Legacy systems are complex not just in structure but in institutional knowledge - the unwritten rules, the workarounds, the political decisions embedded in configurations. Automating discovery and integration may solve the technical problem while leaving the organizational one unresolved.

Still, the trajectory is clear. As AI models become commoditized, the competitive advantage will shift to deployment speed and operational integration. Companies that can roll out agents quickly, iterate on workflows, and scale across business units will pull ahead. June is betting that the path to that advantage runs through automation, not armies of consultants.

The question now is whether enterprises are willing to trust software to do the work that, until recently, required weeks of meetings and a team of specialists camped out in conference rooms.