When Marc Benioff built Salesforce, enterprise software implementations took months and cost millions. Now Benioff is backing a startup that says AI agents can do that same work in a fraction of the time. On August 3, June AI emerged from stealth with a $20 million pre-seed round that includes Benioff’s TIME Ventures, Michael Dell, Diane Greene, Aaron Levie, and George Kurtz — a who’s who of enterprise software history writing a collective check on the belief that AI is about to flip how businesses deploy and maintain the software that runs them.

The Problem No One Talks About

There’s a dirty secret in enterprise AI adoption: the technology is only as useful as the systems it plugs into. And those systems — the SAP deployments, the Salesforce rollouts, the ServiceNow configurations, the Workday integrations — are a mess. Most enterprise software implementations run over budget, over time, and leave behind years of technical debt that becomes a ceiling on what AI can actually do.

Enterprises spend hundreds of billions of dollars each year on this work. Most of it goes to systems integrators and implementation partners who charge by the hour to configure fields, map data flows, write custom code, and maintain integrations that break every time a software vendor pushes an update. This isn’t glamorous work. It’s also not the kind of work anyone puts in a press release. But it’s the reason most AI transformation projects stall before they deliver.

June AI is targeting exactly this gap. Its platform scans a company’s existing systems to understand current business processes, finds bottlenecks, and builds optimized, agent-powered replacements — automatically notifying teams through their existing communication channels. The company integrates across the major enterprise software stacks: Salesforce, ServiceNow, Workday, Oracle, SAP, and Microsoft, with data environment support for Snowflake and Databricks.

The Team Behind It

The founding team isn’t new to this problem. CEO Efrat Rapoport co-founded Bonobo AI, a conversational intelligence company that Salesforce acquired in 2019. After the acquisition, the founding team spent five years inside Salesforce working on AI products — watching firsthand as enterprise customers struggled to connect AI capabilities to the tangled realities of their existing software stack.

CTO Idan Tsitiat, President Barak Goldstein, and Chief Architect Ohad Hen round out the leadership. Between them, the team has built and sold an AI company, scaled products inside one of the world’s largest enterprise software businesses, and watched the gap between what AI promises and what enterprises can actually execute on firsthand.

That experience matters. Implementation problems look simple from the outside and are genuinely hard to solve at scale. The founders have lived that complexity from both sides.

Why the Timing Is Right

Enterprise software implementation has always been painful. What’s changed is that the pain is now blocking AI adoption.

Organizations that want to deploy AI agents — whether for customer service, financial operations, supply chain management, or HR workflows — need those agents to connect to their systems of record. An AI agent that can’t read from Salesforce, write to SAP, or query Workday isn’t useful in most enterprise contexts. And the integrations required to make that connection work are often held together with custom code written by contractors who may no longer work for the company.

The result is a growing category of AI-ready organizations that aren’t actually ready. They have the appetite for AI transformation and the budget to pursue it. What they lack is clean, well-connected infrastructure underneath. June AI is betting that AI agents can now fix that problem faster than human implementation teams can.

The investor lineup supports the thesis. Marc Benioff built a company worth hundreds of billions on the premise that software shouldn’t require an IT department to deploy. Aaron Levie built Box on the idea that enterprise content could work the way consumer software does. Diane Greene built VMware on infrastructure that made complexity manageable. These are people who have spent careers thinking about how to make enterprise technology less painful — and they’re collectively backing the idea that this particular form of pain is now solvable with agents.

What This Means for Business

For business leaders, the June AI launch is a signal more than a product announcement. The signal is this: the friction between AI capability and enterprise deployment is now large enough that a $20 million pre-seed round from some of the most credible names in enterprise software can be raised to address it.

Most organizations still treat implementation as a necessary cost — something to budget for and endure. What June AI is suggesting, and what its investors appear to believe, is that implementation can become a competitive variable. Organizations that can configure, reconfigure, and update their enterprise software systems quickly will be able to deploy AI capabilities faster than competitors still waiting for their systems integrator’s next available slot.

The practical implication is that companies need to think about their existing software stack not as a fixed asset but as infrastructure that either accelerates or constrains their AI roadmap. If your Salesforce configuration hasn’t been updated since 2022, your AI agents are working with stale foundations. If your ServiceNow implementation is held together by manual workarounds, the AI you’re planning to layer on top will inherit those workarounds.

What This Means for Business

If you’re leading a business through AI transformation, the June AI story points to a few questions worth asking now:

What’s actually in your enterprise software stack? Most organizations have a rough sense of what software they use but limited visibility into how those systems are configured, what integrations exist, and where the technical debt lives. Before deploying AI agents on top of your business systems, understanding what those systems actually look like matters.

Who owns implementation when it goes wrong? Enterprise software implementations fail at a high rate. When an AI-powered process fails because the underlying system configuration is incorrect, accountability gets murky quickly. Having a clear owner for implementation quality — not just for the AI layer — is increasingly important.

Is your vendor stack a competitive asset or a liability? The organizations that will move fastest with AI aren’t necessarily the ones with the biggest AI budgets. They’re the ones whose underlying systems are clean, well-integrated, and updated. That’s a different kind of investment than buying the latest model.

Enterprise DNA’s Omni Advisory service helps organizations think through exactly these questions — building AI roadmaps that account for the full stack, not just the AI layer on top. If you’re planning an AI transformation and want to stress-test the foundations before you build, book a discovery call with Sam McKay.

The fact that Benioff, Dell, and Greene are writing checks into a company tackling implementation debt tells you something about where the real bottleneck in enterprise AI actually sits. It’s not the models. It’s the plumbing.