A $20 million startup wants AI to fix its own deployment mess
Getting AI to actually work inside a large company is so painful that a new wave of expensive consultants has emerged just to handle it. June thinks software can replace them.
Key points
- June, an AI deployment startup, raised $20 million in pre-seed funding led by Marc Benioff's Time Ventures in 2025.
- Backers also include Michael Dell and Aaron Levie alongside George Kurtz.
- The four founders previously built and sold Bonobo AI, a voice-to-text company, to Salesforce in 2019.
- Large companies routinely spend weeks stalled trying to connect AI tools to their existing data systems.
- June's platform scans a company's systems, maps the blockages, then builds a step-by-step plan to get AI agents running without outside consultants.
Big companies want AI. Getting it to do anything useful inside their existing systems is a different problem entirely.
That gap has spawned a small industry of forward-deployed engineers, specialists who fly into a company and manually untangle the mess so an AI tool can finally connect to real data and real workflows. Expensive, slow, and, according to one new startup, unnecessary.
What does June actually do?
June scans a company's existing software systems, maps where the blockages are, then builds a plan for deploying AI agents, software programs that carry out multi-step tasks automatically, without outside consultants doing the heavy lifting.
CEO Efrat Rapoport, a former Salesforce executive, co-founded the company with Ohad Hen, Barak Goldstein and Idan Tsitiat. The four built Bonobo AI, which launched a voice-to-text service in 2017 before Salesforce acquired it two years later. After spending years inside Salesforce watching large customers struggle to adopt AI, they decided the problem needed a software fix rather than a staffing one.
The core issue is messiness. A typical large company stores its data across platforms including Salesforce, ServiceNow and Workday, each carrying years of accumulated clutter: duplicate fields, disconnected sources, workflows nobody fully documents. An AI agent dropped into that environment gets confused fast.
"How does an agent know how to operate when you have 10 duplicate database fields that say the same thing, and different teams are using them?" Rapoport told TechCrunch AI.
June's answer is to scan all of that, surface the specific problems, and give teams a clickable list of fixes. Remove these duplicate fields. Connect to this data source. Click build, and June does the work.
Does this actually work for real companies?
Paul Akinmade, chief strategy officer at CMG, a large US mortgage lender, says yes. His team had moved to Claude Code, the AI coding tool from Anthropic, but hit a wall integrating it with Salesforce. He had publicly committed at Salesforce's annual conference to returning with 100 agents running, and weeks of meetings with architects and forward-deployed engineers had produced nothing. June, he says, gave his team a clear view of where to deploy agents and let them move forward safely, even before the two companies held an official kickoff call.
His terms for piloting any tool were direct: "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," he told TechCrunch AI.
That pressure from buyers is exactly the market June is betting on. Our earlier look at what enterprise leaders actually need before deploying AI agents found that most organisations are still measuring the wrong things, which makes the clarity June promises genuinely rare.
What did investors pay for?
The round was led by Time Ventures, Marc Benioff's fund. Michael Dell and CrowdStrike's George Kurtz joined Box's Aaron Levie as co-backers. June did not disclose its valuation.
Rapoport says it came together without much salesmanship: "We didn't even have a deck for this raise."
The real question isn't whether AI deployment is hard. It obviously is. The question is whether a scan-and-fix platform can stay ahead of the consultants who will eventually automate the same moves. Watch for whether June's enterprise customers renew after the first deployment or quietly hire back the engineers they thought they'd replaced.
Watch for: If your company is being pitched AI tools, ask specifically how the vendor handles integration with your existing data systems. The answer tells you whether the AI will work or just sit there.
