Specialization is key in an exploding AI chatbot market

Amid an exploding market for AI chatbots, companies that target their virtual assistants to specialized enterprise sectors may get a firmer foothold than general chatbots, according to Gartner analysts. That’s not news to Zowie, which claims to be the only AI-powered chatbot technology specifically built for ecommerce companies that use customer support to drive sales – no small feat in an industry in which customer service teams answer tens of thousands of repetitive questions daily. Today, the company announced it has secured $14 million in series A funding led by Tiger Global Management, bringing its total to $20 million.

Chatbots — sometimes referred to as virtual assistants — are built on conversational AI platforms (CAIP) that combine speech-based technology, natural language processing and machine learning tools to develop applications for use across many verticals. 

The CAIP market is extremely diverse, both in vendor strategies and in the enterprise needs that vendors target, says Magnus Revang, a Gartner research vice president covering AI, chatbots and virtual assistants.

The CAIP market is comprised of about 2,000 vendors worldwide. As a result, companies looking to implement AI chatbot technology often struggle with vendor selection, according to a 2020 Gartner report coauthored by Revang, “Making Sense of the Chatbot and Conversational AI Platform Market.”

The report points out that In a market where no one CAIP vendor is vastly ahead of the pack, companies will need to select the provider best fit for their current short and midterm needs. 

The secret sauce in the AI chatbot market

That is Zowie’s secret sauce: specialization. Specifically, the company focuses on the needs of ecommerce providers, Maya Schaefer, Zowie’s chief executive officer and cofounder, told VentureBeat. The platform enables brands to improve their customer relationships and start generating revenue from customer service. 

Plenty of other CAIPs provide services for companies that sell products. But their solutions are also targeted to other verticals, such as banking, telecom and insurance.  Examples include Boost AI, Solvvy and Ada. Other chatbots — Ada is an example — can also be geared for use in the financial technology and software-as-a-service industries to answer questions, for instance, about a non- functioning system.

Zowie is built using the company’s automation technology, ‘Zowie X1’, to analyze the meaning and sentiment to find repetitive questions and trends. 

Zowie claims to automate 70% of inquiries ecommerce brands typically receive, such as “where’s my package?” or “how can I change my shipping address?” she says. The solution also includes a suite of tools that allows agents to provide personalized care and product recommendations, she says.

For example, if a customer says, “I would like help choosing new shoes” the system hands the request to a live product expert.

Before implementation, the platform analyzes a customer’s historical chats, frequently asked questions and knowledge base to automatically learn which questions to automate. It uses AI capabilities to analyze new questions and conversations, delivering more automation. 

By analyzing patterns, the AI chatbot can tell when something new or unusual is happening and alerts the customer service team, Schaefer said.

Live agents may also have difficulties upselling customers, so Zowie gives unique insights about customers to agents–what product they are looking at, did they order before and if so, what did they order – and has a direct product catalog integration which enables agents to send product suggestions in the form of a carousel, she added.

Time to optimization

In 2019, Schaefer and cofounder Matt Ciolek developed the all-in-one, modular CAIP designed for building highly customizable chatbots. Schaefer estimates that within six weeks of implementation, Zowie answers about 92% of customer inquiries such as “where’s my package?” and “what are the store hours?”

“Managers and agents don’t have to think about how to improve customer experience — our system will detect new trends and propose how to optimize the system automatically,” she said. “In a way, we automate the automation.”