#USA #Teaneck #Cognizant #Multi-Agent #Neuro AI

Cognizant Launches Open-Source Neuro AI Multi-Agent Accelerator

In a significant move for the field of artificial intelligence, Cognizant (NASDAQ: CTSH) has announced that its Neuro® AI Multi-Agent Accelerator is now available as open-source software targeting research and academic purposes. This innovative resource allows domain experts, researchers, and developers to rapidly prototype and build agent networks for virtually any use case.

The open-source nature of this software is designed to accelerate AI adoption by fostering collaboration in the creation and adaptation of multi-agent systems for adaptive operations and real-time decision-making. Businesses can leverage Cognizant’s Multi-Agent Services Suite to implement these networks on a large scale within commercial environments while effectively managing them under a commercial license.

AI Agents Market Growth

The market for AI agents is expected to experience explosive growth over the next five years, projecting an increase from $5.1 billion in 2024 to a staggering $47.1 billion by 2030. This transition underscores the potential for interconnected agents to unlock new revenue streams and generate scalable business value. Through the development of its open-source Neuro AI Multi-Agent Accelerator, Cognizant is asserting its leadership in AI innovation and its commitment to advancing the capabilities of AI agents.

Notably, industry leaders like Telstra, Australia’s foremost telecommunications and technology company, are partnering with Cognizant to test and implement multi-agent systems. Kim Krogh Andersen, Group Head for Product and Technology at Telstra, remarked, “The open-source release of the Neuro AI Multi-Agent Accelerator will empower our teams to quickly develop prototypes and integrate existing AI agents, thus helping to expedite our software development cycle. We are already witnessing improvements in quality, speed, and efficiency.”

Real-World Applications

Cognizant has recently assisted a healthcare organization in establishing an agent network for contractual negotiations, which has significantly reduced processing times for medical appeals. Additionally, a consumer goods firm has benefitted from better insights into supply chain management through similar agent-driven support. Presently, Cognizant is engaged in over 65 conversations with clients about agent-based AI implementation.

The complexity of building a successful multi-agent network lies in coordinating diverse agents, tools, and knowledge sources, including universal Large Language Models (LLMs) and organization-specific systems such as Service Level Management (SLMs) or Retrieval-Augmented Generation (RAG) frameworks. Cognizant’s Neuro AI Multi-Agent Accelerator aims for seamless integration capabilities with APIs, RAG, and third-party agents like Agentforce from Salesforce, Agentspace from Google, or Crew AI, utilizing its native Model Context Protocol (MCP) or standard API calls. An optional coordination protocol among agents enables self-organization, task delegation, and process management, enhancing efficiency and minimizing errors. The Agent2Agent (A2A) protocol is also supported, promoting inter-agent collaboration across clouds, platforms, and corporate boundaries.

The Need for Experimentation

“Remaining competitive in the age of agent-based AI requires organizations to experiment — to explore how agents can transform business processes and enhance operational efficiency,” stated Babak Hodjat, Vice President of AI Technology at Cognizant. “With the open-source release of the Neuro AI Multi-Agent Accelerator, we’re broadening access to our cutting-edge multi-agent technology, enabling developers to innovate more swiftly and allowing decision-makers, regardless of their technical background, to prototype systems quickly and observe their impacts on key performance indicators.”

Gary Lerhaupt, Vice President of Product Architecture at Salesforce, emphasized the significance of such collaborative efforts: “Agentforce relies on Salesforce’s fully integrated, open, and extensible platform, enabling our ecosystem of partners and developers to drive innovation with trusted AI. Cognizant’s decision to make its Neuro AI Multi-Agent Accelerator open-source stands as an exemplary partnership model that assists our clients in moving swiftly and innovatively.”

Key Features of the Accelerator

• - Intelligent Opportunity Discovery: Input a company name or problem area, and the Agent Network Designer will automatically suggest a tailored agent network aligned with your use case, facilitating rapid ideation and implementation.
• - Rapid and Optimized Customization: Quickly create and modify multi-agent systems using natural language or pre-built templates in areas like lending, customer service, retail optimization, and intranet automation, sharply reducing development times and risks.
• - Scalable, Distributed Operations: Connectors support custom tools, APIs, and third-party agents via the Model Context Protocol (MCP) or standard API calls. A coordination layer enables intelligent self-organization, task distribution, and process management among agents, boosting efficiency and minimizing errors.
• - Secure Private Data: Supports regulated industries like finance and healthcare by isolating confidential information over private data channels to ensure compliance and data protection.
• - LLM and Cloud Provider Independence: Seamlessly switch between open-source and most commercial LLMs as well as public and private cloud providers without system reconfiguration.
• - Extendable Coded Tools: Enhance your agent networks with custom tools essential for making agent decisions based on real-time data or defining logical boundaries that necessitate human intervention as required.
• - Multiple Servers, Distributed Deployment: Run agent subnetworks across multiple servers, enabling scalable architectures that support parallel processing, geographic distribution, or segmented use cases.
• - Data-Driven Network Definition: Completely define and update your agent-based systems via simple configuration files supporting version control, verifiability, and rapid reusability across projects.
• - Agent Testing Functionality: Utilize the Agent Network Tester to identify bottlenecks or failures within your network, providing actionable insights on coordination issues, gaps in agent logic, or integration errors.

Cognizant is scaling agent networks for 330,000 employees through its intranet. Thousands of employees are now utilizing 1Cognizant, an intranet assistant built on the Neuro AI Multi-Agent Accelerator. This tool consolidates and organizes numerous agents to assist employees efficiently with varied tasks, from booking conference rooms and ordering taxis to managing requests like parades or weddings. 1Cognizant can now provide immediate actionable advice and support to employees, enhancing efficiency and breaking down internal silos.

For further information, visit our landing page or check out our blog.

About Cognizant

Cognizant (Nasdaq-100: CTSH) develops modern enterprises by aiding customers in modernizing technology, redesigning processes, and transforming experiences to stay ahead in our rapidly changing world. Together, we enhance daily life.

For more information, visit www.cognizant.com or follow us @cognizant.