Summary
DaveAI's Avatar-powered Chatbot is revolutionizing the customer experience! DaveAI Avatar-powered chatbot mimics a human sales brain, interacts with customers in natural language, understands their requirements & provides hyper-personalized product recommendations in both speech & text: all in real-time. This intelligent voice & chatbot can interact with customers in natural language. The Avatar-powered chatbot discovery improves customer experience across the brand's digital touchpoints by adding a human touch to humanize digital conversations.
Challenge
Customer Product Discovery for all categories starts online. But there is a significant difference between a Sales Person Assisted Discovery Experience Offline v/s Self Assisted Discovery Online. How can a brand provide a personalized experience to its customers on an online platform, by keeping it relevant, interactive & engaging at the same time? DaveAI bridges this gap unlocking a massive opportunity with an Interactive Sales Avatar-powered Chatbot. This is highly relevant for complex industry segments like Automotive, BFSI, Lifestyle Retail & others.
Objective
DaveAI's Avatar-powered Chatbot enables sales scenarios where a complex catalogue is sold to customers or where products are sold to a large number of customers. Our objective was to improve the below-mentioned metrics in order to improve the overall customer experience: 1. DISCOVERY - To enable customers to discover every brand (car/variant/features) 2. NUDGES - To provide Active & Passive Nudges to improve engagement and discovery. 3. FAQs - To provide real-time hyper-personalized answers to Standard Sales Related FAQs – Crowdsourced from MSIL. 4. EVENT ENGAGEMENT - To help users with features like Locating Store, Price List, Downloading Brochure, etc. 5. CONVERSION EVENTS - To lead the customer towards Booking a Showroom Visit, a Test Drive, or a car purchase. 6.RECOMMENDATIONS - To provide Real-time customer product recommendations.
Strategy
The virtual avatar bot was first deployed in 3 varied brands in different channels to get a sense of how customers interacted with the interface as well as the impact it created. This limited live deployment gave us some valuable insights (i) Discovery was a key problem on the web, Discovery was useful for customers who are yet to make a choice, and hence deployment was taken up in the brand pages of the website. (ii) User Surveys were conducted to understand key intents and further develop user stories that are tied to the KPIs of a specific brand. (iii) Conversations were differently designed for each brand as the demographic of the target segment varied as well as their navigation pattern while discovering a car. (iv) Finally, all of these were tied to the overall business KPIs at a brand level and business level. This allowed users to have a seamless experience while browsing a car whereas the KPI of improved sales metrics was met effectively. The bot also integrated with the enterprise’s customer data platform, ensuring that the bot had the context of the user’s previous interactions before interacting with the virtual avatar chatbot. This allowed to create personalized engagement strategies based on the user’s previous state. The AI affinity engine predicts the next best action for users based on their previous interaction, enabling a sustained improvement in KPIs achieved. The virtual avatar chatbot was inspired by the best salespeople of the enterprise. Conversations were designed from insights received from sales teams on ideal user interactions and outcomes. The objective was to provide an experience similar to what a user would receive when interacting with a trained salesperson. One of the key aspects considered is also learning to empathize with the customer’s current state of mind.
Data
Every decision in this avatar-enabled chatbot is data-driven. The first decision to deploy the avatar-enabled chatbot across brands was taken from the results of live 3-month pilots across 3 brands. Today data helps us decide: 1)how the bot performed against KPI targets, More than 5 Million user interactions have been completed with the avatar helping influence sales of >11900 vehicles. 2) how users engage with the bot- What queries were asked, what queries were unanswered? This data helps improve the scope of the conversation and also has the brand teams add more information to the knowledge bank. 3) 19600 hours of engagement data panning across 2.5L unique events were analyzed from Phase 1 to plan the scope of the conversations and intents for Phase 2 and 3. 4) For eg, the entire user interface and experience are being redesigned for each brand based on the engagement events and user behavior analysis. 5) Integrations with 6 different enterprise systems such as CMS, DMS, CDP, Service, Project Management and Product Information Systems were completed during the deployment to enable users to perform various actions in an automated fashion through the avatar chatbot. 6) The AI affinity engine had sufficient engagement data to drive personalization use cases for each customer who engages with the bot This has enabled this virtual avatar-enabled bot, a one-of-its-kind deployment at scale in a car sales scenario. Maruti Suzuki Assistant today is the world’s first AI Virtual Avatar-powered bot that has completed more than 5 million customer conversations.
Solution
Every enterprise sits on tons of sources of truth that can be leveraged during a sales conversation however the sales reps are provided with just a CRM which helps them capture an interaction but does not help them with actionable insights in real-time that can influence a sales conversation. DaveAI does just that - as a Virtual Avatar is a sales companion that enables the sales reps with actionable in real-time while learning from data available across the enterprise to predict what works for which customer. - Dave has the ability to humanize digital conversations with a Virtual Avatar powered bot, that is capable of mimicking a human sales brain. - With a blend of Speech & NLP, 3D Visualisation, and Real-time Recommendations into a single immersive audio-visual experience, DaveAI enables seamless product discovery for customers. - Dave's Empathetic AI platform has the ability to start off with minimal data and learn on the go without the need to have a data scientist. Once there is a certain amount of data, the platform leverages deep neural networks to improve the accuracy of predictions. - Dave's Online learning genetic algorithm Adapts to the sales process of every enterprise & auto-tunes the importance of Specific Products, Attributes, Categories,Trends & New Products & Offers.
Results
1. During the first 3 months: DaveAI's avatar-powered chatbot was used by >2L customers across 3 brand pages where an A/B test was conducted. 50% of users of the brand page were shown the avatar while others interacted with a regular web page. DaveAIs virtual avatar chatbot improved engagement by >50%, and conversion and engagement metrics improved by >20% leading to the decision to have that chatbot for all brands. 2. Present: - Unique Customer Interactions: 18 Million - Leads Qualified: 25,000 + - Dwell Time: 84000 Hours - Loan Applications Influenced: 500,000 - Increase in Average Session Time: 150 % - Avatar helps influence sales of >11900 vehicles. - 26% improvement in Lead Qualification. - 66% Improvement in Engagement Time - 23% Improvement in Engagement Events - 35% Improvement in Conversion Rates - Approx 50% improvement in Cost Saving
