On July 28, 2026, Harmony raised $34 million in seed funding led by Lightspeed Venture Partners to deploy AI agents inside Slack and Microsoft Teams, targeting enterprise IT, HR, finance, legal, and operations workflows. The round included Hitachi Ventures, Fin Capital, Mercer Ventures, and Operator Partners, alongside angels from the founding teams of Wiz and Eon.io. Harmony reports a 70% no-touch resolution rate, meaning seven of every ten employee requests are resolved by AI without a human touching the ticket.
The raise matters for two reasons beyond the dollar figure. First, $34 million at seed is a large commitment for a one-year-old company with about 50 employees. Second, Lightspeed has now backed the same founding team twice. Their first company, Epsagon, sold to Cisco for $500 million. Lightspeed participated in that outcome too.
What Harmony Actually Does
The premise is straightforward: employees spend their workdays inside Slack or Teams. When they need IT help, an HR policy, a software access approval, or a payroll answer, the default experience is to leave those tools, navigate to a separate portal, open a ticket, and wait. Harmony removes that detour.
The platform embeds AI agents directly in the collaboration tools employees are already using. A new hire can type a question in Slack and receive both an answer and a completed action: software access provisioned, policy document surfaced, onboarding task marked complete. The request never enters a human queue unless the agent cannot resolve it.
The technical layer enabling this is what Harmony calls an organizational context graph. The graph connects each employee’s identity, devices, installed applications, current permissions, and work history. When an engineer asks to access a deployment tool, Harmony knows their role, which tools engineers at their seniority level typically have, and what approval chain the request requires. The agent initiates the provisioning, routes the approval, and notifies the employee when access is live. No ticket number. No waiting for a technician to read the queue.
The platform ships with more than 100 prebuilt agents covering IT, HR, finance, procurement, legal, security, and DevOps. Deployment takes days, not months. Critically, Harmony integrates with existing ITSM systems rather than replacing them, so organizations can adopt it without dismantling current infrastructure.
The Resolution Economics
The shift from ticket-based to agent-based service delivery changes resolution economics at every layer of the stack:
| Metric | Traditional Ticketing | AI Agent Resolution (Harmony) |
|---|---|---|
| Average resolution time | 2 to 5 business days | Minutes |
| Employee interface | Web portal or email queue | Slack or Microsoft Teams |
| Routing model | Manual triage, assigned technician | Auto-resolved or escalated with context |
| No-touch rate | 10 to 20 percent | 70 percent (reported) |
| Context available to agent | Ticket history only | Full organizational context graph |
| Deployment timeline | Months | Days with prebuilt agents |
Harmony’s 70% no-touch figure is vendor-reported and has not been independently verified across a broad customer sample. But the directional shift is material even at more conservative real-world rates. An enterprise handling 2,000 IT and HR requests per month at 50% no-touch resolution removes 1,000 requests from the human queue. At an average loaded cost of $15 to $25 per human-handled ticket, that is $15,000 to $25,000 in monthly capacity freed for higher-value work.
The Founders’ Track Record
Nitzan Shapira (CEO) and Ran Ribenzaft (CTO) built Epsagon, a cloud observability startup that made distributed tracing practical for production engineering teams. Cisco acquired Epsagon in 2021 for $500 million. Lightspeed backed that company, and it is backing this one.
Shapira has said publicly that he wants to build a much larger company this time rather than pursue an early acquisition. The $34 million seed, and Harmony’s stated ambition to become a comprehensive Enterprise Service Management platform, signals a longer-duration play than a typical seed-stage company.
The angel investor list adds credibility in a specific direction. Assaf Rappaport, who co-founded Wiz and led it to the largest cybersecurity acquisition in history ($33 billion, Google), is participating. Wiz built its enterprise security platform by embedding into the cloud infrastructure where security teams were already operating, without requiring organizations to replace existing tools. The structural parallel to Harmony’s approach is not coincidental. These are founders who understand how to sell to enterprise IT buyers by reducing friction rather than demanding workflow replacement.
Yoni Cheifetz, the Lightspeed partner leading the Harmony investment, put it directly: “We believe the company is building one of the defining enterprise software platforms of the next decade.” That is an unusually strong public statement for a seed-stage bet, and it reflects how Lightspeed sees the shift from portal-based ITSM to embedded agentic service delivery.
Why This Investment Signals a Structural Shift
The conventional architecture for enterprise service software is portal-first. Build the system of record, then surface a UI on top of it. ServiceNow, Zendesk, Workday, and SAP all follow this model. The employee goes to the software.
The Harmony thesis inverts that assumption. The software comes to the employee, inside the tools they already use every day. This is not a new idea in isolation. Claude Tag brought Anthropic’s AI into Slack as a collaborative agent embedded in organizational workflows. ChatGPT Work extended OpenAI’s agentic capabilities into enterprise workflow automation. What Harmony adds is vertical specificity. It is built for the IT and people operations functions that generate the highest volume of repetitive, structured requests in any enterprise, functions where resolution has historically required human attention at every step.
The broader enterprise AI adoption gap has often come down to integration complexity and change management more than model capability. Employees do not want to learn new tools. IT organizations do not want to replace working infrastructure. Harmony’s design answers both objections: it starts where employees already are and layers on top of systems organizations already operate.
Gartner has projected that by 2026, 40% of enterprise service requests will be fully resolved by AI without human involvement, up from under 10% in 2023. If Harmony’s 70% no-touch rate holds across diverse enterprise environments, the technology is moving significantly ahead of that projection.
Implications for Enterprise Operations Teams
Three practical observations for teams evaluating the agentic service delivery category:
Workflow location beats standalone apps for adoption. Employees resist context-switching to new tools. AI agents deployed where work already happens face substantially lower adoption friction. The highest-leverage deployments are ones that do not require employees to change their behavior at all.
Organizational context is the actual moat. General-purpose AI assistants answer questions. Agents that can take action do so because they know who the employee is, what they are authorized to access, and what the approval chain looks like. The context graph is the infrastructure investment that separates resolution from recommendation.
Measure no-touch rate, not deployment count. An enterprise that deploys 100 agents but resolves 15% of requests without human involvement has done less than one that deploys 10 agents with 70% resolution. The metric that translates to operational leverage is the share of requests that never reach a human queue.
What Comes Next
With $34 million in seed funding, Harmony plans to expand coverage across infrastructure and operations, enterprise applications, cybersecurity, and ERP functions. No customer count or revenue figures were disclosed alongside the announcement. The company operates out of New York and Tel Aviv with approximately 50 employees.
Competitive pressure will come from incumbents. ServiceNow has added generative AI capabilities across its workflow platform. Salesforce Agentforce targets enterprise service automation across sales, service, and IT. Moveworks, a Slack-embedded employee support platform launched in 2019, has operated in this space for years with substantial Kleiner Perkins backing.
What distinguishes Harmony’s entry is timing and team. The agentic AI infrastructure that makes 70% no-touch resolution technically credible in 2026, particularly the context graph and multi-step action capabilities, did not exist at production scale when earlier players launched. The founders have already built and exited one enterprise infrastructure company. And Lightspeed has placed two consecutive bets on their judgment.
For enterprise operations teams still running IT and HR support through email queues and ticketing portals, Harmony’s raise is a concrete signal: the replacement cycle for service delivery infrastructure is underway, and it is arriving inside the tools employees are already using.