Multiagent Systems: Taking Agentic AI to the Next Level in the contact centre
Jurgen Hekkink, Head of Product Marketing at AnywhereNow discusses
Contact centre managers are always looking for cutting-edge solutions to improve their services, balancing operational cost against customer satisfaction. The recent AI revolution and its widespread adoption in the contact centre space have proven that AI can deliver both cost savings and improvements in customer experience simultaneously. This means that implementing this ground-breaking technology is no longer a luxury but a necessity for businesses to stay ahead of their competition and at the forefront of contact centre managers’ minds.
Contact centre operations can be optimised with Agentic AI tools in many ways, such as the automation of routine tasks, knowledge assistants, and smart recommendations, which enhance productivity and improve services. However, to fully utilise the power of Agentic AI, contact centres should look to employ multiagent systems (MAS), whereby multiple Agentic AIs work together as a team to combine their powers and handle more complex scenarios.
What is a Multiagent System?
A multiagent system (MAS) is, in essence, an AI framework where multiple autonomous AI agents interact within an environment to achieve specific objectives. Each of these agents is capable of perceiving its surroundings, making decisions, and taking actions independently. MAS works like a team of specialists in any company, where each member has a specific skill set and responsibilities, working together to achieve a common goal.
Unlike traditional non-AI solutions, which are typically static and predefined, AI agents in a MAS can adapt their behaviour based on real-time feedback and learning. This dynamic interaction enables MAS to handle more complex and unpredictable scenarios, making them perfectly suited to handling customers in a contact centre setting, as dealing with people is inherently unpredictable and requires improvisation.
How to Get Started with a MAS
For decision-makers who want to implement a MAS, knowing where to start can be a daunting challenge, and the fear of stumbling can hold some organisations back. This is especially the case where AI is involved, as the technology is still misunderstood and, in some ways, controversial.
The first step in implementing a MAS is identifying the system’s use case. Knowing the specific areas in the customer interaction process that would benefit from a MAS will provide insight into how best it can be implemented. It is vital to engage with key stakeholders, including contact centre managers, IT teams, and customer service representatives, as their buy-in and support are crucial for successfully implementing a MAS.
After gathering the relevant information, the next step is to create a pilot programme. To ensure the MAS is feasible and worth the resources required, it is best practice to start on a small scale and build up in complexity later. This could be a MAS that handles a specific type of customer inquiry, such as billing questions or technical support, allowing for testing and adjustments before a full-scale rollout.
Another area that must be given considerable attention when planning a MAS is security and compliance. As AI agents often handle sensitive data, it is vital that they are secured and comply with relevant industry standards, as this will provide a strong framework for building a secure and resilient MAS system.
The MAS journey doesn’t end at deployment. Once the system is up and running, it is important to monitor and measure the success of the MAS and optimise it based on this analysis. Regularly reviewing the system’s performance metrics—such as response times, accuracy, and customer satisfaction scores—to identify areas for improvement will ensure that the MAS remains effective and efficient.
Looking to the Future: Current Trends and Developments in MAS
Like most technology, multiagent systems have undergone significant development over the past few years due to the widespread availability and adoption of generative AI (GenAI) and large language models (LLMs). Advancements in machine learning are transforming MAS from knowledge-based information repositories into reactive, action-oriented agents. This also allows MAS to improve over time as algorithms gain access to more data and are tested in real-world scenarios. GenAI has also given rise to the use of autonomous agents that can operate with minimal human interaction.
The best way to utilise MAS is to start in conjunction with skilled employees, where AI and humans can support each other to deliver the best customer experience possible. This approach enables human oversight to mitigate risks associated with GenAI while reaping its benefits. For organisations beginning their AI journey, this strategy offers a safe learning environment to optimise AI utilisation and paves the way for more autonomous applications in the future, balancing innovation with risk management.
As multiagent systems continue to evolve, organisations will see greater benefits to the operational efficiency of their contact centres, as well as enhanced customer experience. It is also important to remember that this will not only benefit the business but also the contact centre agents, who will enjoy a better work environment.
Founded in 2010, AnywhereNow is a Netherlands-headquartered and fast-growing provider of Customer Experience SaaS solutions. AnywhereNow empowers voice and digital dialogues for organisations worldwide and brings to life Agentic AI platforms for increased productivity and effectiveness. AnywhereNow’s products are award-winning, recognised by industry analysts, and trusted by over 2,000 global customers, including Rabobank, DHL, Emirates, KPMG, Swarovski, Mazda, Deloitte, Aldi, Vodafone and Zeiss.
For additional information on AnywhereNow visit their Website
