Ready to reimagine CX? Agentic AI is here.

Early agentic AI deployments are producing significant gains in service speed, interaction success, and even overall brand satisfaction.

But the road to success comes with challenges. How do you ensure your operations are ready? What is the best digital strategy when tech seems in constant flux? And how do you address regulatory and privacy concerns?

Opening here, our new blog series will address these and other questions while sharing real experiences on the ground. 

Welcome to the world of agentic AI.

For many, GenAI is now integral to everyday life. Solutions like ChatGPT, Claude, Gemini, and, more recently, DeepSeek are shaping how people work and interact online.

Now, we are at the dawn of a new ‘agentic AI’ era.

In essence, an AI agent is an entity that operates on a ‘cognitive architecture’: orchestration (instructions) with models and methods (tools, inputs, context), enabling applications and solutions to address defined business needs (goals, objectives). 

AI agents can execute workflows (from simple and specific, to complex or undefined) and drive decision-making autonomously. They continuously learn from interactions, processes and other agents, adapting their behaviors accordingly.

Agentic AI is an important new paradigm for any company looking to transform or enhance customer care with efficiency and scalability. Gartner predicts that within the next four years, at least 15% of daily work decisions will be autonomously made by AI agents – a significant increase from almost 0% currently.

Yet any AI-enabled business must have people at its heart. That is because, while the tech is hugely powerful, long-term success depends on people’s quality of experience. Agentic AI is no exception. A hybrid multi-agentic approach is about integrating artificial and human intelligence seamlessly so customers get the best of all worlds. AI agents empower human agents, complementing their talents while adapting dynamically to customers’ evolving needs in line with the brand. 

So what does a hybrid multi-agent approach look like in customer operations? Broadly speaking:

1. AI as a Virtual Assistant to human customer service/experience agents. Human agents get real-time support and predictive decision-making tools to boost both their performance and productivity. As part of an intelligent experience engine, these enable better targeted and more personalized customer experiences across all channels. Powerful conversational analytics extract new actionable insights from every single customer interaction that human agents can then seamlessly use to make decisions and drive desired outcomes. And they can even benefit from ongoing personalized support and training.
2. Virtual Intelligent Service Agents. Virtual voice and chat agents can be fully integrated for frictionless and personalized multi-channel customer engagement. Customers’ needs can be met quickly, seamlessly and efficiently, freeing up human agents’ time. To assure quality and human-grade experiences, pre-defined mechanisms ensure continuous validation and review of the output (‘human as a supervisor’), re-routing to a human to handle specific decisions or escalations (‘human in the loop').
3. Back-office process automation. Intelligent automation optimizes customer service operations and transforms back-office tasks. Specific workflows can be automated to increase speed and efficiency while cutting costs and eliminating errors. This is usually focused on highly specific capabilities and processes, driving the rise of Vertical AI Agents.

One major advance is the integration of highly specific and context-aware agentic AI applications. This is thanks to Small Language Models (SLMs), described by HFS Research as specialized models designed for specific narrowly defined tasks. These can be trained and fine-tuned for particular domains or contextual challenges, such as specialized terminology, fluctuating demand or the resolution of complex issues.

A key advantage of SLMs is that they can run in less powerful technology environments, enabling deployment closer to customers (such as on edge devices or local servers) where responsiveness and efficiency are critical. What’s more, SLMs can be significantly more cost-effective than other AI solutions, helping to cut the costs of AI deployment - though training and fine-tuning can still require substantial investment.

With continuous data-driven improvement baked in, a hybrid multi-agent ecosystem will fuel ongoing enhancements to customer service and satisfaction, plus operational speed, accuracy and efficiency.

Some solutions even enable business and operations teams to independently configure and deploy AI-powered customer journeys. This can significantly cut implementation costs and timelines while serving to democratize AI technologies, giving business owners more clarity and control.

While many organizations are still early in their agentic AI deployments, we are already seeing notable results in companies we work with, including:

• Significant cuts in agent ramp-up time and typical productivity increases of around 10% 
• Virtual agents handling 50% of retail orders with success rates higher than humans’ 
• Transformed claims processing, commonly reaching up to 30% faster handling times, 90% error reduction, 49% of cost savings in debt recovery, and the list goes on.

However, while the benefits of agentic AI are clear, there are challenges for any business or technology leader to bear in mind – not least how to: 

1. Seamlessly fuse the best of technology and human talents
2. Prepare for next-generation operations, train employees and embed the new operating model
3. Ensure readiness, availability and quality of organizational knowledge and data
4. Ensure all AI outcomes are aligned with brand positioning and value
5. Drive early value while continuously evolving, scaling and adapting the AI.

To address every one of these challenges, a well-defined human-centric AI roadmap is crucial. From here, an evolving hybrid multi-agent strategy will offer a scalable and future-proof route to look after customers and keep the organization at the forefront of the enterprise AI revolution. 

In our upcoming blogs, we will keep exploring how to turn human-centric principles into action, unlocking even greater value for your business. Stay tuned!

-
With the contribution of Luigi Esposito, Head of AI Deployment for EMEA and ESM, and Diana Catalina Velasquez, Head of AI Deployment for LATAM.