A decade of innovation has visibly transformed what retail looks like in 2025. But that pales in comparison to the technological revolution ahead of us. Henri Seroux and Raphaël Hervé discuss the limitless potential of Generative and Agentic AI. We stand on the brink of a revolution unlike anything we have seen before.

“The retail sector today is increasingly complex: high costs, ever-changing consumer behaviour and non-stop technological advances.” This is how Henri Seroux, Senior Vice President EMEA at Manhattan, opened his foreword in the very first Commerce Trends edition ten years ago. That was before the pandemic, inflation reached double digits, and long before the United States imposed new tariffs. Costs have continued to rise and consumer behaviour has changed fundamentally. In recent years, the smartphone has become the device where every buyer’s journey begins, with just a few taps, an order can be placed and paid for and all the brand communication and tracking – post-purchase – can be easily navigated. Thanks to this ease, consumers not only expect, but rather demand, near-instant gratification. Continuous innovation has enabled retailers to keep up in terms of their supply chain commerce strategies. Extensive digitalisation creates far more complexity in business operations, but it also enables a more granular, data-driven, personalised approach too. On the shop floor, every customer can be served with tailored offerings based on their purchase history and individual preferences. Whereas in the supply chain, it’s now possible to define the optimal path for each order – from warehouse or store to the customer – based on cost and preferences. Modern technology makes it all possible.

Revolution

Seroux confirms these developments. “There is a wide variety of products that need to be delivered in different quantities, in different ways, to people who want to receive them at different times – usually faster than before. Fulfilling these needs, often with additional services, normally results in higher costs, while retailers simultaneously face increasing pressure to reduce expenses. These conflicting priorities result in a storm of complex decisions in warehouses and stores that demand lightning-fast, informed, accurate responses.”

The rapid emergence of new technologies like Generative and Agentic AI can help these responses. Most people are now familiar with Generative AI thanks to ChatGPT, but Agentic AI goes further: systems that don’t just respond, but can independently analyse data, prepare decisions, and even execute them if needed. According to Seroux, Generative and Agentic AI mark the beginning of a shift far more profound than the change witnessed over the past decade. “We’re facing a revolution on a scale we’ve not experienced since the birth of the Internet. AI is fundamentally changing how both goods and ideas are produced and used.”

AI for contact centres

Raphaël Hervé agrees. As Senior Director Technical Services, he closely follows developments in AI. “This technology is now enabling solutions that our customers have long envisioned but could never realise due to complexity, time, effort or all three. Take Manhattan Active® Maven, the AI application for contact centres that we introduced in 2024.” Hervé explains that traditional chatbots only provide predefined answers to predefined questions. If a consumer asks something outside the script, the bot fails. “But Maven can respond to more nuanced questions, such as inquiries about shipping costs for a specific order. The application knows which APIs to call to check shipping options, calculate costs, and generate a clear answer. In fact, this was our first use case for Agentic AI, before the phrase was really coined.”

“AI is enabling solutions that our customers have long envisioned but cold never realise.” 

Raphaël Hervé, Senior Director Technical Services & Customer Support at Manhattan

Workforce management

Earlier this year at Momentum, Manhattan’s annual North American customer and partner event, another Agentic AI use case was highlighted, optimising warehouse staffing. A manager walking through the warehouse and unsure about operational progress can ask an AI agent via their smartphone whether all orders will be ready for shipment by the end of the day. The agent replies, sparking a conversation. Why won’t we finish on time? Which zone is short-staffed? Which has spare capacity? Can we shift people to make up for the delay? “These are the questions warehouse managers face every day,” Hervé says. “Previously, answering them required digging through dashboards. Now the AI agent does that, meaning a decision that once took 20 or 30 minutes now takes just a few minutes. And the agent can even implement the decision.”

Building your own agents

Hervé envisions many more use cases, such as using Agentic AI to scan various shipping documents, extract relevant data, and feed it into supply chain execution software. “We’ll see more cases where Agentic AI delivers valuable insights without the user needing to consult dashboards, tables, or logs. Supply chain processes won’t get simpler, but the way we use supply chain software certainly will. This leads to faster, better decisions and lower costs because users spend less time on routine tasks and can focus more on areas where they add real value.”

In addition to the agents Manhattan provides, users can design and deploy their own via Manhattan Agent Foundry. This platform offers tools to customise agents to individual processes and preferences. “Agent Foundry also provides access to all agents, whether self-developed or shared by partners in the Manhattan ecosystem. Users can activate them easily and enable communication between them using standard protocols,” Hervé explains.

Google Agentspace

Manhattan is also exploring making these agents available through Google Agentspace, a market  place for agents that communicate with each other through shared protocols. “I foresee a future where Manhattan agents can talk to agents from other IT systems. That will allow us to automate many processes that today require manual steps and decisions,” Hervé says. “The possibilities are endless. That’s why it’s hard to predict quite what Agentic AI will achieve. Much depends on how ready users are to embrace the technology and adapt their behaviour. That’s what makes this such an exciting development.” Hervé emphasises that Agentic AI won’t be forced on anyone. Companies can move at their own pace. “They can be confident that the agents we release have been rigorously tested. We only launch agents that deliver optimal results. Especially for agents that propose decisions, it’s essential to use reliable data and test thoroughly.”

“AI may not replace people, but it will automate repetitive tasks so they can focus on more meaningful work.” 

Henri Seroux, Senior Vice President EMEA at Manhattan

The right foundation

“The decision to begin developing the Manhattan Active platform pre-2017 has proved especially astute. This cloud-native platform with a microservices and API-first architecture provides a near-perfect foundation for leveraging Generative and Agentic AI. All the effort and over half a billion dollars of investment into research and development is now paying off,” says Seroux. Hervé agrees. “The rise of Generative and Agentic AI confirms that our technological strategy was sound. To apply this type of AI, agents need access to platform data and functionality. That’s possible thanks to the microservices and APIs underlying the Manhattan Active platform.”

Another benefit is that AI can help simplify and accelerate the rollout of order, warehouse, and transportation management systems. “We can speed up implementation without compromising quality. AI can replicate system configurations while respecting local, divisional, or site-specific nuances. A lot of time is wasted on unchallenging tasks like configuring parameters stored in spreadsheets. Agentic AI can dramatically accelerate that process. It may not replace people, but it will automate repetitive tasks so they can focus on more meaningful work,” Seroux explains.

What does the future hold?

Are companies open to Generative and Agentic AI? Both Seroux and Hervé say yes. “We saw this with the adoption of Manhattan Active Maven,” says Hervé. “Beyond that, many people already use this technology in their daily lives. New generations prefer to go to a Generative AI app like ChatGPT. The quality of the answers is improving.” Seroux adds: “The enthusiasm of our customers was also clear from the response to our announcements around Generative and Agentic AI. And what’s great is that no upgrade is required to deploy these agents. With Manhattan Active solutions, you always have the latest version with every innovation, there at your fingertips.”

What will the AI revolution bring? “It would be arrogant to claim we know the answer. The future is unpredictable, especially now. Who could have imagined three years ago that we’d be in the midst of an AI revolution?” says Seroux. “What we do know is that with the Manhattan Active platform, we offer a solid foundation to participate in this revolution and that AI agents will help our software modules work even more seamlessly – from order management to warehouse and transportation management, and from supply chain planning through to overall supply chain execution.”

Raphaël Hervé, Senior Director Technical Services & Customer Support at Manhattan

Henri Seroux, Senior Vice President EMEA at Manhattan
 

Wilt u een fysiek exemplaar van het magazine ontvangen? Stuur een mail naar ce@manh.com o.v.v. Commerce Trends 2025, en vermeld bedrijfsnaam, uw naam, adres en woonplaats. Wij sturen u dan een exemplaar toe.