For the past eighteen months, the business world has been obsessed with "builder" Phase of the generative AI revolution. Enterprises have rushed to deploy autonomous agents to handle everything from customer support to complex codebase refactoring.
However, as these digital workers proliferate, a new, more structural problem has emerged: fragmentation. Agents created in LangChain cannot easily hand off tasks to those created in CrewAI; An agent built into Salesforce has no native way to coordinate with a custom Python script running in a private cloud.
Today, a new startup, BAND (also known as Thenvoi AI Ltd.) quietly emerged with $17 million in seed funding to provide the "interaction infrastructure" These isolated tools need to be turned into a unified and collaborative workforce.
"For agents to become real actors in the global economy, they need ways to communicate, just as humans do," said co-founder and CEO Arick Goomanovsky in an interview with VentureBeat, continuing: "The communication solutions we have today for systems don’t work for agents, because agents are non-deterministic creatures. It’s not just about API integrations."
By introducing a deterministic communication layer that functions as "Slack for agents," BAND aims to move the industry from a collection of fragile experiments to scalable and "agent economy".
Introducing ‘Agent Mesh’
The core of BAND’s thesis is that simply creating and connecting AI agents to human communication tools like Slack causes them to lose context or require constant attention. "rehydration" if they fail and re-enter a conversation.
“You can’t take a group of agents and put them on Slack and expect it to work miraculously." Goomanovsky said.
BAND solves this through a two-layer architecture designed to handle the unique telemetry of AI-AI interaction, the so-called "agent mesh."
This is the "interaction layer" where agent discovery and structured delegation occur. It allows agents to find each other across different clouds and frameworks without requiring developers to write fragile messages. "glue code" for each new connection.
-
Peer collaboration: Unlike existing protocols that are primarily peer-to-peer or client-server, BAND supports full-duplex and multi-peer communication. This allows a group of agents (for example, a scheduling agent, a coding agent, and a QA agent) to work together in a shared environment. "room" with synchronized context.
-
Deterministic routing: In particular, BAND does not use large language models (LLM) to route messages. Using an LLM for routing would introduce the same non-deterministic errors that the platform seeks to resolve. Instead, the platform uses a patent-pending multi-layer architecture to ensure messages reach their destination reliably.
-
The WhatsApp comparison: To handle the anticipated volume of agent traffic, BAND’s infrastructure is based on the same technical stack used by global messaging giants such as WhatsApp and Discord. This ensures that the platform can scale to billions of messages as digital identities begin to outnumber human ones.
If the nesh is the "pipeline," The control plane is "valve". This layer provides the runtime governance that enterprises need before they can safely scale autonomous systems.
-
Authority limits: The platform allows organizations to impose strict rules on which agents can talk to each other and what topics they can discuss.
-
Credential tour: One of the most important obstacles in multi-agent systems is identity. BAND manages how human permissions and security tokens pass from one agent to another. For example, if a human asks Agent A for information and Agent A delegates that task to Agent B, BAND ensures that Agent B only accesses data that the original human can see.
Product, platform and pricing: expanding the multi-agent, multi-model workforce
The BAND product suite is designed to be "frame independent" and "cloud independent," positioning itself as an independent middleware that avoids vendor lock-in. In a market where hyperscalers like OpenAI or Anthropic want companies to stay within their specific ecosystems, BAND offers the flexibility to use the best model across multiple options for the job, including open source and custom, optimized enterprise options.
"No matter where agents run or how they were built, we can group them, allow them to discover each other, delegate tasks, and have full-duplex two-way communication." Goomanovsky said, noting that despite competitive first-party options from model vendors like OpenAI Workspace Agents (announced yesterday) and Agents managed by Claude de Anthropic (announced earlier this month), BAND "They play the role of an independent platform that allows a company to avoid dependence on a supplier.
Currently, the company is seeing the most traction in "advanced technology" sectors, including telecommunications, financial services and cybersecurity.
-
Encoder agents– This is currently the most popular use case. Developers often find that Claude is superior at planning, while Codex is better at code review. BAND allows these agents to work simultaneously, delegating tasks to each other in real time.
-
Customer service and operations: Beyond the code, BAND enables "cross-border" automation. For example, a Workday agent might onboard a new employee, who will then contact a ServiceNow agent to open a team ticket, who will ultimately speak to a purchasing agent to finalize the order.
Understanding the sensitivity of enterprise data, BAND offers three main ways to consume the platform:
-
SaaS: A simple cloud-based platform where agents connect via API.
-
Private/on-premises cloud: The entire platform can be deployed within a customer’s VPC or on-premises to ensure data never leaves your control.
-
The edge: The infrastructure is light enough to be deployed in "flying objects" such as drones (UAV) or even satellites, facilitating communication between agents in physically isolated environments.
Early adopters of BAND (and enterprises in general) are already mixing and matching model-driven AI agents from various vendors, so it seems time to provide a global solution.
As Goomanovsky put it: “Advanced developers don’t use a single coding agent. They realize that Claude is very good at planning, Codex is much better at review, and today there is no way to create that two-way interaction between coding, review, and planning agents. We allow it.”
Licensing, governance and pricing
BAND operates as a commercial entity, focusing on providing "enterprise level" stability and security. While the platform integrates with open source frameworks such as LangChain and CrewAI, its own core routing and control technology is proprietary and patent pending.
For enterprise IT leaders, the "Control plane" It’s less about communication and more about auditability. BAND provides full observability on every agent interaction, creating a transcript and "paper trail" for autonomous actions.
This is a "complementary" solution to existing railing products; While a barrier can protect a single agent from rapid injection, BAND protects the the whole system of cascading failures caused by one agent misinforming another.
The company launched a tiered pricing model designed to appeal to everyone from individuals "agent enthusiasts" to global corporations:
-
Free ($0/month): Designed for individuals. It allows up to 10 remote agents and 50 active chat rooms, although it only retains data for 24 hours.
-
Professional ($17.99/month): Aimed at startups and growing R&D teams. This tier increases the limits to 40 agents and 250 active chat rooms with email support.
-
Company (custom): Offers unlimited agents, custom data retention policies to meet compliance requirements, and full API access to BAND. "Memory API".
Towards the ‘universal orchestrator’
The emergence of BAND coincides with a shift in the way analysts view the AI market. Gartner has predicted that by 2029, 90% of companies implementing multiple agents will require what they call a "Universal Orchestrator". Similarly, Forrester has recognized the "Agent control plane" as a distinct and emerging market category.
The company was founded by Goomanovsky and Vlad Luzin, who combined their expertise in Israeli intelligence, cybersecurity, and multi-agent systems to build BAND.
Goomanovsky sees the platform not only as a tool, but also as a fundamental layer for the next Internet era.
"Communication is the most fundamental problem in computing," -Goomanovsky observed. "When new beings emerge, the first thing they need is a way to communicate with each other… We are the internet agent".
The $17 million seed round was led by Sierra Ventures, Hetz Venturesand team8. Sierra Ventures’ Tim Guleri emphasized that BAND is building the "missing layer" that makes large-scale collaboration practical.
This capital will be used to expand the engineering team and accelerate the development of the "design partner" ecosystem, which already includes the main North American telecommunications companies and European digital payments companies.
As agents move from digital novelties to key drivers of enterprise workflows, "glue code" What holds them together will become the most critical piece of the stack. The launch of BAND marks the first serious attempt to standardize that glue, which is chaotic. "band" of agents in a synchronized and governed symphony.
