For the last eighteen months, the corporate world has been dealing with this topic "builder" stage of the generative AI revolution. Enterprises have raced to deploy autonomous agents to handle everything from customer support to complex codebase refactoring.
However, as these digital workers proliferated, a new, more structural problem emerged: fragmentation. Agents built on LangChain cannot easily task those built on CrewAI; An agent embedded in Salesforce has no native way to associate a person with a custom-built Python script running in the cloud.
A new startup today, BAND (also known as Thenvoi AI Ltd.) came out of stealth with $17 million in Seed funding to secure. "interoperability infrastructure" these isolated tools must be transformed into a unified, collaborative workforce.
"To become a real player in the global economy, agents need ways to communicate like humans." In an interview with VentureBeat, co-founder and CEO Arick Goomanovski continued, "The communication solutions we use for systems today do not work for agents because agents are non-deterministic entities. It’s not just about API integrations."
by introducing a deterministic communication layer that acts as "Space for agents," BAND aims to move the industry from a fragile set of experiments to a scalable one. "agent economy".
Introducing ‘Agent Mesh’
The crux of BAND’s thesis is that simply creating AI agents and plugging them into human communication tools like Slack causes them to lose context or require constant work. "rehydration" if it fails and they re-enter the conversation.
“You can’t take a bunch of agents and put them on Slack and expect it to magically work," Guomanovski said.
BAND solves this through a two-layer architecture designed to handle the unique telemetry of AI-to-AI interactions. "agent mesh."
Here it is "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 brittle "sticky code" for each new connection.
-
Multilateral cooperation: Unlike existing protocols, which are primarily peer-to-peer or client-server, BAND supports full-duplex, multi-peer communication. It allows a group of agents (such as a planning agent, a coding agent, and a QA agent) to work together on a common task. "room" with a synchronized context.
-
Deterministic routing: It should be noted that BAND does not use Large Language Models (LLM) to route messages. Using LLM for routing will introduce the same non-deterministic errors that the platform is trying to address. Instead, the platform uses a patented multi-layer architecture to ensure that messages reach their destination reliably.
-
WhatsApp comparison: To handle the expected volume of agent traffic, BAND’s infrastructure is built on the same technical stack used by global messaging giants like WhatsApp and Discord. This enables the platform to scale to billions of messages as digital identities begin to outnumber people.
If Nash "pipes," It is the Control Plane "valve". This layer provides the runtime management that enterprises require before they can safely scale autonomous systems.
-
Limits of Authority: The platform allows organizations to enforce strict rules about what agents can talk to each other and what topics they can discuss.
-
Passage of power of attorney: One of the most important bottlenecks in multi-agent systems is identity. BAND controls how human permissions and security tokens are passed from agent to agent. For example, if a person requests information from Agent A and Agent A delegates this task to Agent B, BAND ensures that Agent B only has access to the information that the original person is allowed to see.
Product, platform and price: scaling the multi-agent, multi-model workforce
Designed to be a product suite of BAND "framework-agnostic" and "cloud-agnostic," positions itself as a standalone middleware that prevents vendor lock-in. In a market where hyperscalers like OpenAI or Anthropic want enterprises to stay within their own proprietary ecosystems, BAND offers the flexibility to use the best-for-business model across multiple options, including open source and fine-tuned, custom enterprise options.
"No matter where agents work or how they are configured, we can connect them, let them discover each other, delegate tasks, and have full-duplex, two-way communication." According to Goomanovsky, despite the competition, as the first part options from model providers OpenAI’s workspace agents (announced yesterday) and Anthropic’s Claude managed agents (announced earlier this month), BAND "acts as an independent platform that allows the enterprise to avoid vendor lock-in.”
The company is currently seeing the most traction "technologically advanced" sectors, including telecommunications, financial services and cyber security.
-
Coding agents: This is currently the most popular use case. Developers often find that Claude excels at planning and Codex excels at code review. BAND allows these agents to work simultaneously, delegating tasks to each other in real time.
-
Customer Support and Operations: Besides code, BAND enables "cross border" automation. For example, a new employee might be hired by a Workday agent, who then contacts a ServiceNow agent to open a ticket for the equipment, and finally talks to a purchasing agent to finalize the order.
Understanding the sensitivity of enterprise data, BAND offers three main ways to use the platform:
-
SaaS: A simple cloud-based platform where agents communicate via API.
-
Private Cloud/On-Premises: The entire platform can be hosted in a customer’s VPC or on-premises environment to ensure data never leaves their control.
-
The Edge: The infrastructure is light enough to deploy "flying objects" facilitate communication between agents in physically isolated environments, such as drones (UAVs) or even satellites.
Already early adopters of BAND, and enterprises more broadly, are mixing and matching AI agents powered by models from different providers, so the time has come to introduce a comprehensive solution.
As Goomanovski says: “Advanced developers don’t use a single coding agent. They understand that Claude is very good at scheduling, Codex is better at reviewing, and today there’s no way to create a bidirectional interaction between coding, reviewing, and scheduling agents. We enable that.”
Licensing, administration and pricing
BAND operates as a commercial entity and focuses on providing "enterprise grade" stability and security. Although the platform integrates with open source frameworks such as LangChain and CrewAI, its core routing and control technology is proprietary and patent pending.
For enterprise IT leaders "Control plane" It’s less about communication and more about verifiability. BAND enables complete observation of each agent interaction by generating transcripts "paper trail" for autonomous actions.
This is a "complementary" solution of existing protective products; While a protective shield can protect an agent from a quick injection, a BAND does the whole system from successive failures caused by one agent misinforming another.
The company launched a tiered pricing model designed to capture everyone individually "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 stores data for 24 hours.
-
Pro ($17.99/month): Aimed at startups and growing R&D teams. This level increases the limits to 40 agents and 250 active chat rooms with email support.
-
Enterprise (Individual): Unlimited agents offer individual data retention policies and full API access to BAND to meet compliance requirements "Memory APIs".
Towards a “universal orchestra”.
The emergence of BAND coincides with a change in analysts’ view of the artificial intelligence market. Gartner predicted By 2029, 90% of enterprises deploying multiple agents will require what they call "Universal orchestra". Similarly, Forrester recognized "Agent Control Plane" as a distinct and emerging market category.
The company was founded by Goomanovski and Vlad Luzin, who combined their backgrounds in Israeli intelligence, cyber security and multi-agent systems to create BAND.
Goomanovski sees the platform not just as a tool, but as the foundation layer for the next era of the Internet.
"Communication is the most basic problem in computing," Guomanovski noted. "When new entities emerge, the first thing they need is a way to talk to each other… We are an Internet agent".
Led a $17 million Seed round Sierra Ventures, Hetz Venturesand Team 8. Sierra Ventures’ Tim Guleri emphasized that BAND is building "missing layer" this makes large-scale collaboration practical.
This capital will be used to expand the engineering team and accelerate development "design partner" an ecosystem that already includes leading North American telecommunications companies and European digital payment companies.
As agents become a key driver of enterprise workflows from digital innovation "sticky code" will be the most critical piece of the stack that holds them together. The launch of BAND is the first serious attempt to standardize this glue, making it chaotic "band" agents becoming a synchronized, controlled symphony.
