Imagine a digital bot in a showroom or bank who understands your preferences and helps you make the right pick. DaveAI, a Bengaluru based startup takes customer experience to an all-new level with such interactive bots. The firm, powered by the Intel Startup Program, works in retail, automobile, and banking, financial services, and insurance (BFSI)
In a store, you have salespeople who introduce the products or services the firm has to offer and influence you. This is missing in a virtual sales experience. In product segments like auto, banking, or lifestyle retail, such product introduction has moved online first but enterprises still rely on traditional e-commerce, which is not designed for complex product segments.
DaveAI is a digital human avatar that can assist customers with everything that is done in an offline store. It can demo products, showcase products in 3D, help visualise various configurations, and also nudge customers with relevant recommendations. The avatar can talk to customers in different languages and can be deployed on any preferred channel such as kiosks, websites, or immersive environments like the metaverse.
With Intel’s backing, DaveAI is delivering this experience at the edge. In retail, these are in-store kiosks, and in banks these are AI Micro branches. AI Micro branch offering replaces a physical branch and enables users to interact with an AI avatar in a language of their choice to conduct banking transactions.
“DaveAI avatars are powered by vision, speech, visualisation, and an AI sales brain. It’s a complete package and not just a good looking digital persona,” says Sriram PH, one of the co-founders of DaveAI.
The avatars can learn on the go based on customer interactions, unlike traditional business intelligence or analytics systems. In some cases, the avatar is set up with very minimal data for the system to learn and improve as more and more customers interact. DaveAI uses a multi-dimensional affinity engine that creates an affinity graph in real time to map customer preferences vs product attributes. The online learning algorithm adapts these affinity scores based on real-time interactions.
With all intelligence combining physical and digital spaces, Sriram says that delivering high performance and impactful experiences on low graphic processing unit (GPU) is a big challenge, which they are solving with the help of a proprietary technology. “The avatar for every interaction leverages compute and graphic-heavy AI workloads, such as computer vision, speech and NLP, 3D visualisation, and a flexible data model. To create a low latency and realistic experience at each touch point is a technology challenge that we solve. It is a constantly evolving space that requires to stay ahead in the innovation curve as the ecosystem evolves,” he adds.
All of these workloads are running at edge right now and allow the system to learn in real time, which means it needs to be a high-power compute. How to optimise these complex workloads at the edge is an ongoing challenge. With the open alliance partnership with Intel, the firm is working on joint research projects to improve these experiences.
Currently, DaveAI is a full-stack platform, which is available as a subscription model for customers.
The platform is also available on a perpetual license model in case the enterprise mandates deployment on their cloud or premises. Their partner model allows partners to leverage the platform subscription and further deploy the same to customers with required customisation and configuration. The firm plans to expand and is open for partnerships with system integrators and services partners and those with experience in global markets as implementation partners and value added resellers.
