Customer Service Gets a Facelift Courtesy of AI

Whether they know it or not, consumers are presented with AI applications throughout their day. Maybe they weren’t aware of that fraud attempt turned away by their credit card company. And maybe they didn’t think of the chatbot that helped them return a holiday gift as AI. Are consumers ready to converse with a visual representation of a live person?

A Toronto-based company called Meeranda says they are. This week, it will be part of the Microsoft Reactor accelerator program, one of just 10 companies invited to participate. CEO Raji Wahidy claims his company will be the first to market with its visual AI application, helped in no small measure by inputs received and relationships formed at several such events since the company was founded early last year.

“When we first started out everyone thought we were building a chatbot,” Wahidy says. “No, we’re not building a chatbot. And there was the impression that we are out there to compete with ChatGPT. It’s not that either. A lot of people didn’t realize that we are more focused on the video generation aspect of the solution, depicting a human who literally looks like you and I, responding back to you in real time.”

Here’s how it works. Suppose a restaurant chain with 15 locations wanted to adopt an AI-driven interface for support and customer communications. If the chain worked with Meeranda, it would connect through its API and design a facsimile of a human face to represent its brand, spokespeople, and, if necessary, its contact center. The chain would work with Meeranda to train its proprietary algorithm to include all the information needed to perform basic functions such as chat, customer service, and even reading a menu. Once the basics are up and running, natural language processing allows it to learn details about the customer it is interacting with. That learning about the customer stays with the Meeranda AI. Unlike basic chatbots, Meeranda can appear to be the same person when and if a consumer revisits a website or mobile app and will remember recent interactions.

In fact, suppose a large multinational corporation has a spokesperson that it wants to become the face of its AI-generated face. Meeranda can design and train the AI spokesperson to sound, act and look like the spokesperson.

Wahidy is designing Meeranda to serve three different use cases: sales, support and marketing. To use the restaurant chain example for sales, a customer could visit the company’s website and ask to hear tonight’s specials. Meeranda will read them, and when the customer returns to place an order, it will remember its details as well as its previous interactions. It could even demo a dish being prepared. Then for support, imagine that the customer has a gift card that couldn’t be redeemed or had a bad experience. Meeranda will sense different levels of support needed. At Tier One it can potentially resolve simple requests. If support needs get more complicated it can kick the case to a human-centric contact center.

“Meeranda can actually work with the organization to conduct co-marketing campaigns,” Wahidy says. “So, you want to promote your products, you want to promote your services, you want to share certain discounts or promotions with your clients. Meeranda can now do that on your behalf.”

While still in Beta mode Meeranda has been tested by companies of all sizes, including the banking and automotive sectors. It recently concluded participation as part of Amazon’s AWS Build Accelerator, which is a 10-week program designed to guide strategic decisions on product development. In addition, through Microsoft for Startups Founders Hub, AWS Activate, Google for Startups, as well as OVHcloud for Startups, it has secured just under $1 million in funding to develop its product and fine tune its strategy.

That strategy has been informed by data. Before developing the product, market research from IPSOS, Forrester Consulting and Clutch found that 77% of consumers surveyed found corporate chatbots frustrating while 88% wanted to speak with a live agent yet were even more frustrated with the long wait times to reach one. The research also revealed that startups tend to hire their support, marketing and sales staff too early, whereas global multinationals tend to hire too many. Lastly, the quality of service received through outsourcing was found to be lacking, according to Wahidy. This all has led to 30% of consumers abandoning a brand due to poor customer experience and 73% cancelling their ongoing orders altogether.

The strategy has also been informed by Wahidy’s prior experience as an engineer at Ericsson and global VP of business operations at Vodaphone. Wahidy also has experience at a startup that used AI tools to help treat marginalized psychiatric populations in prisons and other institutions. The sum total of his experience, he believes, has prepared him to answer the question posed at the beginning of this article. Are consumers ready to talk to an AI?

“Consumers will understand that they are talking to someone who looks like them but is not a human,” he says. “Look to how people have learned to act in other virtual environments. They recognize their environment, and they act accordingly. With AI I think consumers will realize that by interacting with it you’re eliminating a lot of the problems waiting for a long time for service, and sometimes getting unsatisfactory service. But you won’t get the wrong response from an AI, and you’ll never get a bad attitude because they are trained to be on all the time. We humans have bad days, right? We all have bad days. But an AI never had a bad day.”