REZO.ai: The Conversation Builder
Aprajita Sharma September 16, 2019
Manish and Rashi Gupta. Both did Integrated M. Tech in Mathematics and Computing from IIT-Delhi while Rashi also holds a PhD in Biometrics from the University of Helsinki. Before launching their start-up, Manish worked as the chief technology officer at RateGain that specialises in hospitality and travel technology solutions. Rashi was a data scientist at WNS Global, a business process management company.
2) The Trigger
Compared to global standards, customer care in India is quite appalling, and just like others, the Guptas had bad service experience. People are often stuck with automated response systems or 'inane' chatbots or have to repeat their queries and grievances to customer support executives before a resolution can be reached. Every business must address this pain point to prevent customer churn, but bringing in more resources is not a scalable solution and could be a financial drag on companies. So, the duo decided to develop a conversational AI engine to deliver real-time responses to customer queries 24x7 with minimum human intervention.
3) How It Works
When working with a brand, the start-up obtains customer conversations up to six months and feeds the historical data to its AI engine so that its machine-learning algorithms can identify and learn multiple intents (more than one issue featured in a single query), the vocabulary used and the right responses which should be given out. Once a customised solution is developed, Rezo integrates the same with the brand's WhatsApp, Facebook and e-mail interfaces and also with its chatbot software to automate the entire customer relation management (CRM) process.
Besides machine learning, the start-up uses a host of other technologies, including natural language processing, predictive analytics and feedback mechanism, for this purpose. The automated workflow helps brands understand their customers' concerns, reduce wait time and also grow revenues by pushing cross-sell and up-sell via these interfaces. Rezo claims it has cut down handovers (complicated queries forwarded to customer support) by 70 per cent and increased efficiency by 45-70 per cent.
4) Revenue and Growth
The company has a performance-based revenue model and only charges for queries which do not get escalated to executives. "If we are not adding value, we do not charge," says Manish. Rezo is looking to raise funds in the next three months and bring on board at least 50 brands within a year.