Marc Benioff-Backed June Raises $20M to Fix Enterprise AI Deployment

Companies are spending heavily on artificial intelligence, but turning promising AI demonstrations into dependable business systems remains a stubborn challenge. A new startup called June believes the solution is to use AI not only inside enterprise workflows, but also to handle much of the complicated implementation work required to deploy those systems.

June emerged from stealth on August 3, 2026, with $20 million in pre-seed funding led by Marc Benioff’s Time Ventures. The unusually large early-stage round also attracted backing from Dell Technologies founder Michael Dell, Box CEO Aaron Levie and CrowdStrike CEO George Kurtz. The company did not disclose its valuation.

The startup was founded by Efrat Rapoport, Ohad Hen, Barak Goldstein and Idan Tsitiat, a team with previous experience building and deploying enterprise AI products. Their central argument is that creating an AI agent is becoming relatively easy, while connecting that agent to the fragmented databases, legacy platforms and approval processes inside a large company remains difficult and expensive.

June Wants to Automate the Work Around AI Implementation

Large enterprises rarely operate through a single, clean software environment. Customer information may be stored in Salesforce, employee records in Workday, IT processes in ServiceNow and analytics workloads in Databricks. Different departments may also maintain duplicate fields, conflicting naming systems and years of accumulated technical debt.

These complications can prevent otherwise capable AI agents from functioning reliably.

June’s platform is designed to scan a company’s existing systems, identify how its business processes currently work and detect the technical or operational bottlenecks that could block an AI deployment. It then creates an implementation roadmap showing which databases need to be connected, which duplicate fields should be removed and which workflows need to be redesigned.

According to Rapoport, users can then approve individual tasks and allow June to begin building or configuring parts of the required system directly inside the organisation. The platform can also notify relevant teams through their existing workplace communication channels.

This approach effectively treats AI deployment itself as an automation problem.

Instead of requiring a large team of consultants or engineers to study every system manually, June wants an AI platform to map the organisation, recommend the necessary changes and help execute them.

Enterprise AI Spending Is Rising Faster Than Successful Deployment

The timing of June’s launch reflects a widening gap between enthusiasm for generative AI and the number of projects producing measurable business results.

Menlo Ventures estimated that enterprise spending on generative AI reached $37 billion in 2025, up from $11.5 billion in 2024. That represents growth of more than three times in a single year. About $19 billion of the 2025 total went to applications and software built on top of underlying AI models.

Yet much of this investment has not progressed beyond experimentation.

ISG’s 2025 enterprise adoption study found that only 31% of the AI use cases it examined had reached full production. Companies had spent an average of approximately $1.3 million on AI initiatives, but only one in four projects was delivering the expected return from business growth. Around half were producing the efficiency improvements organisations originally anticipated.

McKinsey’s 2025 global survey similarly found that moving from pilots to company-wide impact remained unfinished work for most organisations. The consultancy reported that fewer than one-third of respondents were following most of the recommended practices needed to adopt and scale generative AI successfully.

These figures support June’s view that access to capable models is no longer the only major obstacle. Companies also need clean data, redesigned processes, integration with established software, clear ownership and safeguards for how AI-generated decisions are used.

Why Forward-Deployed Engineers Have Become So Important

The difficulty of integrating AI into established companies has led to growing demand for forward-deployed engineers, commonly known as FDEs.

These specialists work closely with customers to adapt AI products to their internal data, tools and business requirements. Rather than handing over standard software and leaving the customer to configure it, an FDE may spend weeks or months inside the organisation solving integration problems and customising the technology.

The model has become particularly relevant in enterprise AI because each company’s data architecture and operational processes are different.

However, it can also make deployments expensive and difficult to scale. A software company may be able to sell the same AI model to hundreds of customers, but it cannot easily provide a large implementation team to every account.

Rapoport described this as a paradox: AI is expected to reduce manual work, yet deploying it frequently increases demand for consultants, architects and professional-services teams. June is attempting to convert some of that human-intensive implementation process into a repeatable software product.

The company says its platform could complement consultants and FDEs rather than eliminate them entirely. In practice, however, customers may be attracted to June precisely because it could reduce their dependence on specialised deployment teams.

The Founders Previously Sold Bonobo AI to Salesforce

June’s founders have already built and sold an enterprise-focused AI company together.

Their previous startup, Bonobo AI, launched a voice-to-text and conversational intelligence service in 2017. Salesforce acquired the company in 2019, before the current generative AI boom transformed the enterprise software market.

Following the acquisition, the founders spent several years working on Salesforce’s artificial intelligence initiatives. Rapoport later served as vice president and head of Salesforce’s research and development operations in Israel.

The team’s time inside Salesforce exposed it to the implementation problems faced by large customers. Even when companies had access to advanced AI products, their existing systems were often too fragmented or poorly structured to support dependable automation.

The founders eventually left Salesforce and reunited to build June around that problem.

Investor confidence in the team was reportedly strong enough that the founders raised the $20 million round without preparing a formal pitch deck, according to Rapoport.

A Mortgage Lender Became an Early Test Case

June has already tested its platform with CMG Financial, a major US mortgage lender.

CMG chief strategy officer Paul Akinmade said the company had moved its software engineering work to Anthropic’s Claude Code relatively quickly. Integrating AI agents into Salesforce proved considerably more difficult.

Akinmade had publicly committed to returning to a future Salesforce conference with 100 operational agents, but his team reportedly spent weeks consulting architects and forward-deployed engineers without resolving its deployment obstacles.

June’s software gave CMG a clearer view of where agents could be introduced and what changes were required to deploy them safely. Akinmade said he had specifically warned the founders that he did not want another product that depended on specialist engineers or operated as a difficult-to-understand black box.

The early example is significant because it demonstrates the market June is targeting. The startup is not primarily trying to persuade companies to experiment with AI. It is targeting organisations that have already selected models and created agent prototypes but cannot connect them reliably to real business processes.

The Agent Market Is Growing, but So Are Concerns About Failure

June is entering the market as enterprise software companies aggressively promote AI agents capable of completing tasks across sales, finance, customer service, human resources and IT.

Gartner has predicted that task-specific AI agents will become increasingly common across enterprise applications and that agent ecosystems will eventually work across multiple business functions and software platforms. The research firm expects a third of user experiences to shift from traditional application interfaces toward agent-based front ends by 2028.

However, the firm has also warned that more than 40% of agentic AI projects could be abandoned by the end of 2027 because of high costs, unclear business value and immature technology. Gartner estimated that only around 130 vendors offered genuinely agentic products, despite thousands of companies using the term in their marketing.

That tension creates an opportunity for implementation-focused companies.

June’s biggest challenge will be proving that an AI system can accurately interpret the complexity of a large organisation without introducing new risks. Enterprise databases often contain sensitive customer, financial and employee information, while automated changes to business processes can have consequences that are difficult to reverse.

Customers will therefore need transparency into June’s recommendations, strong access controls and reliable human approval mechanisms.

Deployment Could Become the Next Major Enterprise AI Market

The first phase of the generative AI race centred on developing increasingly powerful foundation models. The next phase is becoming a contest over who can make those models useful inside real organisations.

Deloitte’s 2026 enterprise AI research found that employee access to AI increased by 50% during 2025. Companies also expected the share of organisations with at least 40% of their AI projects in production to double within six months, showing that pressure to move beyond experimentation is intensifying.

June is betting that enterprises will not solve this transition simply by purchasing another chatbot or agent builder. They will need technology capable of understanding the complicated systems underneath those tools.

The company’s $20 million pre-seed round gives it substantial resources to develop that platform, but its long-term position will depend on whether it can deliver repeatable deployments across organisations with very different data structures and compliance requirements.

For now, June represents a broader shift in the AI market. As foundation models become more accessible, competitive advantage is moving away from simply possessing an AI model and toward successfully connecting that model to the systems where companies actually conduct their business.