Conversational AI technology is becoming increasingly popular in the sales industry. AI-powered conversational interfaces are already being used by companies like Google, Facebook, and Amazon, as well as companies like Intercom, which provide a conversational AI solution to a vast amount of startups. One startup that caught our attention in terms of delivering added value, as well as a significant increase in ROI for its clients, is Rep, an AI-Powered Sales Associate that has been able to produce a 250% conversation rate uplift as well as a 17% revenue uplift for its clients in the e-commerce space. Rep’s AI uses behavioral analysis to detect when customers are about to abandon a website; then, they offer them contextual and personalized shopping assistance in real-time. We had a chance to talk to Yoav Oz, Rep’s Co-Founder & CEO, and ask him several questions.
Hi Yoav, with other conversational AI products and services (like Intercom) on the market, what makes Rep different? How did you and your team make sure it provides truly personalized experiences and not just generic ones?
Rep was built specifically for e-commerce websites and was designed to tackle e-commerce funnel problems. We developed our propriety NLP engine called CNLU (Commerce-related Natural Language Understanding), which was trained to understand shopping-related conversations between a consumer and a brand. By addressing problems in the e-commerce funnel, our solution adapts itself in the form of conversation to where the customer is in the shopping journey (At the homepage vs. viewing a product vs. Abandoning the cart)
The e-commerce space has been seeing significant innovation recently with decentralized solutions like Bolt, which are able to provide a smoother one-click checkout experience; this is following several analyses indicating almost a 70% cross-industry shopping cart abandonment rate. What steps did Rep take to make sure your conversational AI solution helps improve the numbers for consumer brands? What did you take into consideration to make the AI Sales associate more contextual in specific stages of the funnel?
We’ve discovered that the biggest issue is actually at steps of the funnel before the checkout experience. For example, roughly 40%-60% of the visitors never get to view a single product or add any product to their cart. So by optimizing the first step of the e-commerce funnel, the homepage, we were able to show a significant uplift in sales straight away.
The idea is to let the customers shop on their own until they face a certain issue where we can help, whether it is being overwhelmed by the vast amount of product options available, not being able to select between a few products, or viewing a product that they like but not being sure about it.
For that cause, we developed what we call an “AI drop-off detection engine.” This machine-learning-based component identifies disengaged customers who are predicted not to continue to the next funnel step in order to offer contextual and personalized shopping assistance.
Beyond that, our solution adapts itself to every step of the funnel based on the customer journey on the website, segmentation, and purchase history. As a result, a customer approached by our solution at the homepage will get a completely different experience than a customer approached by our solution after viewing a specific product. For example, if a customer viewed product X, the conversation will most likely start with something like: “Hi there, not sure about product X? Can I answer any questions about it, or can I show you other items?”
Some consumers do not like buying online because they feel it lacks customization, assistance, and direct conversation. They also claim that current sales solutions feel too robotic, generic, and time-consuming. Do you think there will ever be a conversational AI service that really compensates for the lack of direct human touch?
This is exactly what we are developing, and this is our vision for the future. By inserting AI into the equation, we not only know when is the optimal moment to approach a customer but can understand what type of conversation would yield the best results in terms of conversion and simply provide an upgraded customer experience.
By using behavioral analysis, our solution knows what product categories the customers are interested in and which products they spent more time viewing. Based on that knowledge, it completely adapts the conversation in the right direction, which is just what a brick-and-mortar sales associate would do. When we brainstorm over new features and product improvements, we always try to imagine: what would retail sales associates do if they had superpowers? If they could see everything, know everything, and be available 24/7.
Obviously, you can’t share it all, but what are some things that are currently in Rep’s product roadmap?
We think big, and we want to power any consumer brand with its own branded AI-powered sales associate. We plan to do so on any digital platform, whether it’s Facebook, Instagram, Whatsapp, or creating a branded sales associate avatar that would help shoppers on the Metaverse.