Marc Benioff Backs $20 Million Startup That Sells AI to Fix Companies' AI Problems

A startup called June came out of stealth Monday with $20 million in pre-seed funding and a pitch that cuts against the industry's own sales copy: AI hasn't made deploying AI easier. It's made the mess bigger.

The round was led by Marc Benioff's Time Ventures, with backing from Dell Technologies founder Michael Dell, Box CEO Aaron Levie, and CrowdStrike CEO George Kurtz, according to TechCrunch. June declined to disclose its valuation.

The company's four founders — CEO Efrat Rapoport, Ohad Hen, Barak Goldstein, and Idan Tsitiat — aren't newcomers. They previously built Bonobo AI, a voice-to-text company that predated the current transformer-model boom. Salesforce bought Bonobo in 2019, and the team spent years afterward building AI tools inside Salesforce before leaving to start June.

Rapoport, a former Salesforce executive, told TechCrunch that companies trying to bring AI agents into daily operations keep hitting the same wall: their own data is a disaster. "You have fragmented data across these platforms. You have complex workflows. You have years of technical debt," she said.

The Fix Nobody Wants to Admit They Need

Here's the part that should embarrass the AI industry a little. The standard answer to "AI doesn't work in our company" has been to hire more people — specifically, so-called forward-deployed engineers who parachute into a business to hand-wire AI tools into legacy systems like Salesforce, ServiceNow, Workday, and Databricks.

Rapoport calls that out directly: "The industry's answer to AI implementation is, let's hire more and more and more people." That approach relies on staffing rather than automation.

June's bet is that an AI system can do what those engineers do. It would audit a company's existing tech stack, spot duplicate database fields, flag broken workflows, and hand teams a step-by-step build plan through their own communication channels. "Building an agent template is the easy part," Rapoport said. "The hard part is getting it to work with the mess underneath."

The real-world example TechCrunch cited involves Paul Akinmade, chief strategy officer at CMG, a major U.S. mortgage lender. Akinmade moved his company's software engineering work onto Anthropic's Claude Code quickly, but ran into problems trying to integrate it with existing systems. That's exactly the kind of friction June says it exists to solve.

Why Investors Skipped the Deck

Rapoport told TechCrunch the founders didn't even need a pitch deck to close the $20 million round. Pre-seed rounds with no deck and no disclosed valuation are common in Silicon Valley when investors are betting on a founding team's prior track record rather than a proven product. Bonobo AI's acquisition by Salesforce and the team's subsequent years building AI inside a Fortune 500 tech giant clearly did the persuading here, not a functioning product with paying customers at scale.

June is entering a crowded and unproven category. Forward-deployed engineer firms and AI-integration consultancies have multiplied over the past two years precisely because the problem is real and nobody has cracked it cheaply. If an AI tool could reliably untangle a company's Salesforce-ServiceNow-Workday spaghetti without a small army of human consultants, it would meaningfully affect corporate IT budgets. It would also put a lot of those forward-deployed engineering jobs at risk, jobs that exist ironically because AI adoption created more complexity rather than less.

The company remains at an early stage. It has no announced customers, no case studies beyond a CMG anecdote about a different product (Claude Code), and no independent verification that its "scan and roadmap" approach actually works at enterprise scale. Every AI infrastructure startup since 2023 has claimed it will finally make agents reliable inside legacy systems. Most of that promise remains unproven in public.

June says it's now working to sign enterprise customers and build out its automated audit-and-deploy platform. Whether Fortune 500 companies actually hand a startup access to their messiest internal data systems — the fragmented fields, the technical debt, the years of workflow rot Rapoport describes — is the open question that will decide if this $20 million bet pays off or joins the long list of AI middleware companies that raised money on a strong founding story and little else.