Details
The client, a European shared mobility services provider with a huge customer base, was struggling to deal with a massive number of customer queries which were of similar nature. The queries, generally related to their service offerings- charges, reservations, terms & conditions, subscription plans, etc. took a lot of time of service department for resolution, despite being redundant in nature.
The scenario was leading to deterioration in service quality as customers with complex queries had to wait for long for query resolution. To bring efficiencies to client servicing project, the client collaborated with Intellica.AI to create an automated FAQ Bot that would handle tier-1 queries to reduce overhead on customer service executives.
RASA Chatbot
- As the volume of queries was large, there were a host of challenges associated with the development of an automated FAQ Bot that could service repetitive queries without a glitch. Some of the challenges involved:
- Lack of a proper system to gauge the effectiveness of the sponsorship campaigns
- Inadequate data and analytics capabilities to measure brand exposure quantitatively
- Lack of a parameter to compare the brand's exposure over other competitors
Solutions
Intellica.AI conducted a thorough analysis of the type of incoming queries to identify and categorize common topics and queries. After deep research, a pool of 61 unique queries was identified which were structured into 14 broad topics. To bring down the cost of operations, RASA Stack was chosen as it is an open source platform.
Other frameworks charge on the basis of the request volume which would escalate the overall cost. In this case, the client was only required to invest in infrastructure without thinking about the transaction cost. Some other innovations and functionalities introduced in the solution were:
- Use of supervised learning approach and development of personalized training statements
- Development of custom actions via web hooks to get dynamic information from the data store
- Extraction of custom and named entities to capture domain-specific data related to the customer queries
- Creation of personalized stories for contextual conversation management
- Incorporation of an industry-centric spell check features to ensure inclusion of business-specific terminologies
- Support for 2 different languages making the FAQ Bot a multilingual solution with better connect
- Use of JavaScript Components to deploy FAQ Bot on website targeted at the client’s users
- The Fallback mechanism for handing over complex queries with customer details to human customer service executive.
- Owing to the introduction of automated FAQ Bot, the client was able to achieve higher efficiency and the following benefits:
- Savings in time and effort spent on managing redundant customer queries
- Cost-efficient handling and management of the large volume of common queries owing to open source framework
- Seamless end-user experience in case of handover to the human executive because of a full-proof fall back mechanism
- Detailed analytics on the number of queries received, responded and fallbacks
- Insightful analytics about the handover of queries for future upgrades