Tribal, a Tel Aviv-based startup founded in 2025, is betting that enterprises need a dedicated layer to deploy AI agents inside their existing core systems, not another bolt-on dashboard. The company raised $10 million in a seed round led by Team8, with participation from DYDX Capital and angel investors including former Salesforce Israel R&D head Efi Cohen. The 15-person team, built around Salesforce, Wix, and Spot.io veterans, already counts ADAMA, WalkMe, and Dot Compliance as customers. Read the Tribal funding announcement<\/a> for the full breakdown.<\/p>\n The interesting signal here isn’t the fundraise, it’s the customer list. Tribal is a company that founded itself in 2025 and already has named enterprise customers deploying in production. That speed is only possible if enterprises are no longer evaluating whether to run AI agents inside their systems and are now deciding how to do it without breaking compliance, security, or existing architecture. If you’re still in evaluation mode, these customers are already running laps.<\/p>\n Tribal’s core claim is that AI agents fail in enterprise environments not because the models are weak but because the agents don’t understand organizational context, meaning the access permissions, approval chains, department-specific data structures, and operational constraints that make a company actually function. The approach, deeply embedding context before deployment rather than bolting it on afterward, is the right instinct. The MCP (Model Context Protocol, a standard for connecting AI models to the data and tools they need to act) framing from CEO Yoav Kolodner signals Tribal is building to an emerging architecture standard, not a proprietary lock-in play. That’s a credible long-term bet, though a $10 million seed round against the infrastructure ambitions Kolodner describes means execution risk is the dominant variable right now.<\/p>\n The vendor category Tribal is staking out, call it contextual deployment infrastructure, is going to get crowded fast. Salesforce, ServiceNow, and Microsoft are all pushing agent frameworks through their existing enterprise footprints, and they have the distribution Tribal doesn’t. Tribal’s real window is the two to three years before those platforms mature enough to satisfy the customers who need agents working across heterogeneous system stacks today, not in a roadmap. The question your team should already be weighing is whether your current AI deployment vendor has a credible answer to cross-system organizational context, or whether you’re funding a gap you haven’t formally named yet.<\/p>\n MCP is an emerging open standard, think of it as a universal adapter, that defines how AI models connect to external tools, databases, and systems in a structured, permission-aware way. It exists because AI agents that can’t reliably understand what they’re allowed to access, and what the data actually means in context, produce errors that are dangerous at enterprise scale. Tribal’s bet that MCP adoption is accelerating is the architectural premise the entire product roadmap rests on.<\/p>\nWhat this means for your business<\/h2>\n
Concept deep-dive: Model Context Protocol (MCP)<\/h2>\n