Manhattan Associates have announced two new generative AI solutions aimed at empowering customers using their Supply Chain Commerce platform solutions. The first is Manhattan Active Maven for customer service. The solution leverages generative AI to deliver an enhanced chatbot experience to customers. The second is Manhattan Assist, a generative AI-powered assistant that offers support within its product suite.
Ellie Crawford, Director of Product Management for Manhattan Associates, commented, “GenAI is one of the most exciting technologies we’ve seen in years, and it promises to redefine customer service experiences.
“Manhattan Active Maven and Manhattan Assist represent an exciting application of this technology in supply chain and commerce, delivering new levels of personalisation, productivity, and cost-savings.”
Manhattan Active Maven
The first generation of chatbots was limited in functionality and responses. Manhattan Active Maven represents a new generation of chatbots that leverages LLMs to provide a wider context for answers, empathize with customers, and provide a more human touch.
The new chatbots are trained on Manhattan Active Omni’s rich commerce functions. This enables the chatbot to answer a wider range of enquiries. No matter what the complexity of the questions according to Manhattan. Customers can ask for information about order changes, cancellations, returns, exchanges, store locations, product information and more. Active Maven interprets the request and responds, dynamically adapting the conversation to meet a customer’s actual need.
Active Maven provides the first layer of support for customer service teams. Thus freeing up their time to respond to more in-depth requests that require further action. Importantly, the Active Maven chatbots operate 24×7, can answer enquiries no matter the time of day. They can also scale to provide answers to commonly asked questions when widespread issues occur.
The new chatbots can also respond appropriately to the tone and language used by the caller. They will assess whether the language used is positive, neutral, or negative. Then respond appropriately to help calm a situation if necessary or to retain the positive vibe.
The solution is simple to implement. It just needs embedding in the website and provides a first line of support. Where necessary, the chatbot understands when and how to transition the handover to a human agent if further action is required. That handover includes a detailed history of conversations. It also records, summarises and updates customer records. So subsequent human interactions have easy access to all conversations held with customers. What isn’t clear is the pricing of Manhattan Active Maven.
Manhattan Assist
The second new solution is embedded within its solutions. Providing a valuable help agent for those looking to leverage the Manhattan suite of solutions. Manhattan Assist understands the product functionality, API Structures and other information that administrators and developers need to carry out both common and uncommon tasks.
Again, using natural language, users can request information about how the product works and also how the deployment is configured. Manhattan Assist is included in all Manhattan Active Solution subscriptions. In theory, it should help lighten the load on the Manhattan customer services teams in the same way that Manhattan Active Maven does for its customers’ service teams.
Enterprise Times: What does this mean
Manhatta Associates has now embedded generative AI into its solutions. It isn’t clear whether Manhattan Maven is included for customers. Hopefully, it is, and customers can quickly adopt it and see the benefits. Manhattan Assist will prove equally useful for both customers and Manhattan Associates themselves. Introducing generative AI onto its platform to provide support is a sensible move. Surprisingly, few firms have done this.
What is missing from this announcement and the information available on the Manhattan Associates website is more detail. Both in terms of technical detail about the AI applications themselves but also about Manhattan Assist. Its potential seems large. At the moment, Manhattan Assist seems to hold information and answer questions about the product and how it works.
Looking forward, could it go a step further and provide insights about a configuration? Would it be possible to identify dormant accounts, for example, or where a configuration isn’t optimal? The next step is using it for programmatic or prescriptive analysis.
