NTT DATA and Microsoft simplify multiagent AI systems at scale

While some organizations are still focused on integrating single-purpose GenAI solutions into their workflows, their forward-thinking peers are moving on to the exciting challenge of managing multiagent systems at scale.

These systems involve multiple AI agents, each with specialized abilities, working together to achieve complex tasks. They collaborate to achieve their goal by sharing information and dividing tasks in a flexible, adaptive way.

In this way, they are helping organizations to innovate, modernize their operations and deliver great customer experiences.

Smart technology, complex challenges

However, integrating agents across diverse technology stacks from different vendors and platforms is a complex undertaking. This can lead to significant governance, risk management, interoperability, data-exchange and security issues.

Another challenge in managing multiagent AI systems is keeping all agents operating within regulatory and compliance frameworks while also managing the operational aspects of these systems, such as task allocation and workflow management. This can be resource-intensive.

Protecting multiagent systems from cyberthreats and ensuring data privacy are also significant concerns.

Making multiagent management easy

NTT DATA is already developing, implementing and maintaining advanced AI and GenAI solutions around the world for clients in several industries.

We were therefore familiar with the challenges of orchestrating agentic AI services that involve different agents and vendors, even as the industry embraced developments aimed at creating a common language for agents, like Anthropic’s Model Context Protocol (MCP) and Google’s Agent2Agent (A2A) Protocol.

Now, by combining our Agentic AI Services for Hyperscaler AI Technologies with Microsoft’s multi-agent capabilities added to Azure AI Foundry Agent Service, we have found a way to observe and orchestrate agentic AI solutions across platforms and offer these capabilities as an innovative, high-value managed service. 

NTT DATA is leading the way as the first global systems integrator to develop a working prototype using support for multi-agent workflows in Azure AI Foundry Agent Service.

Through this partnership with Microsoft, we enable multiagent systems to operate seamlessly, adapt to changing environments and deliver secure, efficient solutions.

We chose to work with the Microsoft platform because it supports the integration of agents from various vendors and platforms, meaning that they can now work together effectively.

It also enables the creation of adaptive workflows that can adjust to the specific needs of each task, which supports operational efficiency.

And, lastly, the platform includes robust security and compliance tools to protect multiagent systems and ensure they meet regulatory requirements.

• ALSO READ → A force for good: How AI and GenAI are reshaping our world

A multiagent system in action

One of the most compelling examples of NTT DATA and Microsoft’s collaboration is a multiagent ticket management system for IT service management, just launched at Microsoft Build, the annual conference for software engineers and web developers.

Our system streamlines and automates the process of handling tickets, such as support requests or incident reports. It deploys five types of AI agents:

• Data-gathering agent: Collects data from users.
• Ticket-classification agent: Classifies and opens tickets.
• Prioritization agent: Checks and prioritizes tickets based on their nature.
• Routing agent: Decides whether to close the ticket, route it to a human or send it to a resolver agent.
• Resolver agent: Uses MCP to resolve and close certain issues.

Tickets are routed to the most appropriate AI agent to resolve issues quickly and effectively. Agents collaborate in real time, delivering fast, high-quality responses, and the system can automatically adjust workflows based on the requirements of each ticket.

The business value is significant: The system cuts response times while raising resolution rates and overall customer satisfaction by managing tickets more effectively and efficiently.

This model showcases how flexible and scalable our approach is, as it allows the integration and management of agents across platforms and hyperscalers. It also supports the addition of specialized agents, like guardian agents for security or red-teaming agents for compliance — all updated regularly as part of our end-to-end managed service.

The ability to build and manage agents in a client’s environment while monitoring them remotely is also crucial.

AI agents for insurance

We also developed a conversational multiagent system to support personalized, high-value customer interactions for an insurance company.

In this system, different AI agents focus on specific customer requests — for life-insurance details, general policy information and the company’s health and wellness program, among others — or act in a supervisory capacity. To deliver a seamless, efficient service, all these agents work in harmony.

AI agents working within multiagent systems are set to revolutionize processes and outcomes in a range of industries — for example, product design and optimization in manufacturing, product recommendations in retail, loan processing in banking, treatment plans in healthcare and emergency response systems in the public sector.

The power of Microsoft

Microsoft’s multi-agent workflows in Foundry Agent Service and Semantic Kernel are essential tools in the orchestration and monitoring of multiagent systems like these.

The multi-agent workflows allow for the creation, management and observation of complex interactions between different agents. Semantic Kernel, on the other hand, helps in externalizing data, making it easier to integrate and manage information across various agents and platforms.

Combine these technologies with NTT DATA’s deep global expertise in systems integration and achieving business goals, and you have a winning recipe for multiagent AI success.

• ALSO READ → Meet your smart new coworkers: Microsoft Azure AI agents

The business impact

“NTT DATA is leading the way as the first global systems integrator to develop a working prototype using support for multi-agent workflows in Azure AI Foundry Agent Service. Their ability to build, manage and orchestrate agentic AI models across multiple platforms is a game changer, enabling cross-platform observability, simplifying complex multiagent deployments and accelerating client ROI,” says Yina Arenas, Vice President of Product and Core AI at Microsoft.

When you simplify multiagent complexity and manage these systems at scale with the help of NTT DATA, you benefit from:

• Simplified and automated business processes: Streamlined workflows and automated processes reduce operational overhead and improve efficiency.
• Better operational outcomes: Improved governance, risk management and security lead to better operational outcomes.
• Cost savings: Efficient management of multiagent systems results in cost savings and increased ROI.

Take the next step

NTT DATA and Microsoft are making multiagent AI simple by collaborating to address the challenges of integration, governance, operational efficiency and security.

We’re setting a new standard for multiagent system management. Contact us to see how this will benefit your organization, too.