Cognitive computing: A diverse and fast-growing market

Growth in AI software platforms from vendors such as IBM, Google, AWS, and Microsoft is increasing at a rapid rate, according to David Schubmehl, research director for cognitive/AI systems at IDC. However, developers are using a variety of approaches. “A lot of organizations are developing their own AI-enabled applications from open source frameworks, and we are also seeing increasing use of commercial AI-enabled solutions,” he noted.

Virtual sales assistants expedite customer engagement

The applications for cognitive computing and AI are numerous and heterogeneous. Among the more mature applications are virtual sales assistants and chatbots, which are already in use on many websites. The Los Angeles Film School, based in Hollywood, Calif., which offers associates’ and bachelors’ degrees fields related to the entertainment industry, is one example of an organization that is using an AI-powered sales assistant.

The website for the school is generally the first contact point for prospective students, who fill out an online form to request additional information. The school was getting hundreds of leads per month from these requests. However, the follow-up from its salespeople was not meeting the target, which was to make seven attempts to connect in the course of the first week after an initial contact.

Seeking a greater degree of consistency and persistence, the marketing department decided to try an AI-powered solution. After exploring options, the school selected the AI Sales Assistant from Conversica, which it discovered at a trade show. The choice was based primarily on the maturity of Conversica’s product. Conversica provides virtual assistants for sales and customer success in a SaaS environment. The solutions integrate with sales and marketing products such as Marketo, HubSpot, and Salesforce.

Conversica assisted the marketing department at the Los Angeles Film School in setting up the virtual assistant, a process that took less than 2 weeks. Under the new system, leads are still routed to salespeople first, and the initial attempt at contact is still made by them. However, if the salesperson does not reach the target of connecting within seven attempts over 7 days, the lead is passed along to “Stephanie.”

The virtual assistant interacts with prospective students in a way that is difficult to distinguish from that of a human salesperson, asking questions and providing relevant follow-up to the answers received. When the prospective student confirms interest and contact information, the lead is then passed back to a human salesperson. Interactions can occur either via emails or texts.

The outcome of launching the virtual sales assistant was an increase of 30%–50% over the previous 2%–3% conversion rate. The school attributed the full percentage increase to the Conversica virtual sales assistant. In addition, because there is a record of each interaction, whether via the salesperson or the virtual assistant, sales executives can follow up as needed. The system allows a considerable level of accountability at each level.

“Conversica’s products are not designed to replace human salespeople, but to complement them,” said Alex Terry, CEO of Conversica. “The transition from salesperson to Conversica and back again is very smooth and natural.” As for the human-sounding nature of the assistant, “Every month, people who are interacting with our assistants ask them out on dates or try to hire them,” Terry commented. “This is a good indicator that they do not come across as robotic.”

Conversica has designed solutions for seven industry verticals and has additional ones in the pipeline. “After having provided this service for nearly a decade, we have about 500 million messages through our conversational AI platform, and our systems have been trained on millions of real-world conversations,” continued Terry. Each customer receives a unique version tailored to its particular needs. Some conversations are similar across the applications, even across different verticals, and others are selected or created during implementation. These become part of the conversation library in Conversica’s solution.

“We always have humans in the background ready to jump in if the AI does not know what to do next,” explained Terry. “With some systems, customers know they are stuck in a loop and get frustrated. There is actually a fairly narrow happy path for customers engaged with virtual sales assistants, so it is important to keep the engagement positive.” If the customer goes in an unexpected direction, the virtual assistant hands over the interaction in real time to a human who keeps the conversation going.