Currently programmed with natural language processing (NLP), chatbots are trained to perform within a specific domain — meaning they are not capable of going off topic if a user were to share that they are sick, depressed and so on.
With brands looking to make stronger connections with customers, technology is being developed to move chatbots in a new direction that will make conversations less transactional and confined to predetermined responses, and instead, more personalized. This could result in humans turning to chatbots for emotional support, and in turn, brands being able to create a more personalized experience through each interaction and conversation.
To better understand the future of chatbots and how their technology is shifting to be able to handle deep human conversations, Agustin Huerta, VP of Technology at Globant. Huerta explains how chatbots are being programmed to identify keywords and respond to deep level conversations.
Digital Journal: How common are chatbots becoming for businesses?
Agustin Huerta: Due to advancements in AI, chatbots are able to solve customers’ problems, personalize messages and improve the interaction between a brand and its customer, all of which contribute to adoption amongst businesses. We see chatbots being used most frequently for customer support inquiries, given customers want a quick response when they reach out to an agent. By nature of the technology, chatbots are able to meet this need faster than a human since they are available 24/7 and as such, free up humans from more mundane tasks so that they can focus on higher value add contributions.
DJ: What advantages are businesses seeking to leverage from chatbots?
Huerta: Businesses are largely looking to chatbots to enhance customer experience. Chatbots give them an advantage here by supporting human reps and alleviating common pain points such as long wait times or having to reroute calls. At the end of the day, customer satisfaction is king and thanks to the speed, 24/7 access, advanced learning tech and natural language processing, chatbots can deliver an exceptional customer experience every time. With an exceptional customer experience tends to come greater loyalty and revenue, which makes these investments a priority for businesses.
DJ: Generally, how do customers react to chatbots?
Huerta: Generally, they react well. Sometimes they don’t even realize they’re interacting with a chatbot. However, if a customer has a question they need a quick answer to, they know a chatbot can help them find an answer no matter the time of day or queue in the customer service support line.
Frustrations arise when chatbots aren’t able to identify what exactly a customer is asking or looking for from the support center – which are heightened if the chatbot doesn’t respond in an expedited manner. Overall, a customer’s reaction to chatbots has to do with whether or not the chatbot accurately identifies their problem and provides a solution that helps them reach a resolution in a timely manner.
DJ: What are the main limitations with chatbots?
Huerta: The limitations with chatbots have to do with the models they operate under. Most chatbots are retrieval-based, meaning they spit out responses from a predefined bank. If brands or businesses are looking to deliver a more personalized experience for customers, chatbot’s responses need to mirror the language and diction of a human. This is possible with generative-based model chatbots since they use natural language processing (NLP) and take the context of the conversation into account when responding and offering a solution.
DJ: How are chatbots set to change in the future?
Huerta: Chatbots are still working on becoming better at natural language processing. As they work to become more advanced, chatbots will play a larger role in building customer loyalty as brands are able to leverage them to personalize messages to customers, check on their satisfaction with a product and provide solutions to any problems at record speed, all of which play an essential role in creating a positive customer experience. With generative-based models becoming a more common development with AI, chatbots will be able to produce more organic, authentic responses that closely mirror conversations that one would have with a human – given that they take the context of the conversation into consideration more closely when determining a response.
DJ: What development specifics are needed to take chatbots to the next level?
Huerta: The model (retrieval based or generative, as of now) of chatbots will need to evolve in order to take them to the next level. As previously mentioned, most chatbots follow a retrieval based model. If chatbots are to be taken to the next level, more work needs to be done in regard to the generative model of the bots in order to be able to handle deeper conversations and mimic the diction and tone of conversation that a human provides.
DJ: Will consumers react positively to more ‘emotionally tuned’ chatbots?
Huerta: That will depend on the consumer and the level of support they’re seeking. For those looking for emotional support, then yes, I anticipate more empathetic chatbots will garner a positive reaction. If they’re looking for a deeper, more emotional level of support (healthcare, for example) compared to those who are inquiring about the status of a product they ordered, responses that are pre-generated won’t fill their needs. Emotionally-attuned chatbots (made possible with generative-based models) will be able to do this to an extent through NLP and thus satisfy those who are looking for support. Time will tell how successful this will be.
DJ: Will future chatbots do away with the human customer services function completely
Huerta: No, chatbots will never completely replace humans. There will always be those scenarios where a human’s emotional and intelligent response is necessary depending on the context of the conversation. In the healthcare industry, as an example, a customer could easily message a chatbot and begin divulging deeply personal information about their physical well-being that the technology isn’t capable of processing. In these cases, human customer service representatives would need to intervene.
While I do believe chatbots will play a supporting role in delivering high-quality customer service, they’ll never completely do away with humans, since they aren’t capable of addressing every scenario. In their current state, chatbots are just one more agent within a contact center and should be managed as such. Humans are still necessary when it comes to meeting complex customer demands.
