Data science and artificial intelligence development platform company Zerve AI Ltd. today announced a new multi-agent system engineered specifically for enterprise-grade data and AI development.

The new system, part of the Zerve 2.0 release, transforms AI agents from simple code assistants into fully active collaborators, according to the compay. Zerve Agents can now participate across the entire development lifecycle — planning, provisioning infrastructure, building and deploying data and AI products at scale.

With the new system, teams can activate multiple AI agents on any problem using natural language prompts. The agents will present a plan and then get to work, generating new Zerve canvases, creating and connecting code blocks, writing code, orchestrating infrastructure and automating various parts of the data workflow.

The agents are part of the Zerve Operating System, which connects directly to internal infrastructure, whether cloud-based or on-premises, to provide a visual, code-native canvas for human and agent collaboration. The platform allows for seamless integration into existing development pipelines with full compute control and access to code, data and workflows.

One key feature of Zerve 2.0 is “The Fleet,” a distributed compute engine that enables massively parallel code execution using serverless technology. It is particularly suited for large-scale operations such as making multiple concurrent calls to large language models, offering speed and efficiency without driving up costs.

Another addition is App Builder, a feature that allows data and AI teams to turn their work into scalable products without needing front-end or DevOps support. The builder enables natural language querying within apps by integrating Zerve Agents directly into the user interface.

Zerve Agents are designed to mirror human developers by learning from the data and adapting when encountering errors or failed experiments. The context-aware approach helps accelerate development timelines and reduces the need for manual intervention in debugging and troubleshooting, the company claims.

“If you’ve tried out ‘vibe-coding’ agents and wished you could apply them to real enterprise AI and data workflows, now you can,” said co-founder and Chief Executive Phily Hayes. “Zerve is hosted in your own environment, with your own SDLC policies, connected to your LLMs, data repositories, single sign-on and the whole shebang, giving your team and AI agents a secure, productive environment in which to explore, build, test and deliver data and AI products in record time.”

Zerve is a venture capital-backed startup that has raised $11.4 million in funding, including a seed round of $7.6 million in October. Investors in the company include Paladin Capital Group, Elkstone and Rob Hickey, former executive vice president of DataRobot.

Image: SiliconANGLE/Reve