CXOToday has engaged in an exclusive interview with Mr. Durgesh Choudhary, CPO Convin
- How can Artificial Intelligence for Customer Support assist Businesses?
A few years back, predicting what customers wanted, what they liked or didn’t like, and what would lead to customer retention was a big hurdle for businesses. Considering that around 60-70% rely on solving their problem on phone calls, support agents spent prolonged hours on these calls without concluding with a practical resolution. Handle times were nearly 10-15 minutes, and companies struggled with 40% or less FCR rates.
Today, it is easier to guess that refund policy discussions often lead to negative sentiment, while discussing early bird discounts excites the listener.
AI has been instrumental in improving customer satisfaction by making the customer support experience faster and more effective. With the introduction of AI-based chatbots, up to 70 percent of call, chat, and email inquiries have been reduced, which clearly opens up more space for agents to handle personalized and tricky queries.
AI-backed tools fix agent challenges with automated coaching and an updated knowledge database. Support agents can make improvements in ongoing customer calls with AI-backed real-time analytics. It is easier to create automated query-handling processes where human intervention is needed but at a later stage.
In my opinion, AI has blessed the customer support landscape with the 3 P’s–Proactivity, Personalization, and Predictability.
2. How is conversation AI strengthening Customer Support in BFSI?
Conversation AI works on multiple support channels–such as calls, chats, and emails– preferred by customers. The conversations are recorded, transcribed, and analyzed for insights by Machine Learning algorithms and Natural Language Processing. The insights and transcriptions strengthen the customer support teams in all industries.
In the BFSI industry, we’ve noticed that the nature of queries is critical, time-sensitive, and confidential. Customer support teams need to handle these queries carefully(stay compliant) and offer faster resolutions.
Conversation AI stages various analyses and assistance to agents through conversation behavior analysis and customer intelligence. There’s more transparency into common objections, customer sentiments, frequently asked queries, emerging customer trends, and competition in BFSI. Visibility into customer sentiments and emotions improved, which helps in accurately understanding the financial choices and expectations of today’s digital customers.
Moreover, unlike earlier times, agents in the BFSI sectors can offer better resolutions to customers with real-time guidance and analytics. Agents are more prepared and handle complex questions far more efficiently.
3. How can the BFSI industry leverage AI to elevate CSAT and sales conversion?
At Convin, our customers realize a 27% improvement in CSAT scores and a 21% increase in closure rates. Honestly, don’t expect AI to make it happen overnight. Achieving the transformation in customer experience and sales conversions takes time because AI automates processes that impact the business parameters. Agent-friendly AI solutions offer actionable recommendations that retain existing customers.
4. How is AI impacting compliance monitoring in BFSI?
One core issue in BSFI is adhering to regulations and compliance. Agents often get carried away by sales targets and make false commitments. We have encountered several cases of mis-selling policy benefits and coverage in the auto insurance industry. As per IRDAI, 35,178 complaints of mis-selling cases were reported in 2019-20. Bank agents often disclose half the information to clients and leave out the fine print.
Two AI-based solutions have successfully improved compliance monitoring- Automated QA reduces random sampling of calls and manual auditing, ensuring every call is screened for compliance breaches. Not just screening but Auto QA scores agents based on their performance and indicates improvement areas. Once compliance violations are identified, agents can receive Automated Coaching sessions presenting the top performer in the improvement area.
5. Development and adoption of voice and conversation AI systems.
Voice AI capabilities operate on ML(Machine learning) and NLP(Natural language processing) algorithms. Speech is converted to text and analyzed using machine learning algorithms. Algorithms are pre-defined and trained to handle similar data sets.
The development and adoption of conversation AI have evolved from the following transformations in Voice AI technology:
- Complexity of language– Demand for unpredictable conversations, multiple languages, multiple accents, and code-mixing needed faster processing and better algorithms. The new Voice AI technology makes it necessary to build advanced algorithms supporting complex language, code-switching, domain-specific models, noise filtering, faster analysis, and actionable data.
- Real-time data analysis and usage– In the past, post-activity analysis delivered valuable insights, but it was a reactive style of processing data. Data processing power is improving, enabling real-time analysis and proactive decision-making (reducing processing time dramatically).
- Accuracy of analysis- For an analysis to be accurate, the text data must be accurate. The speech-to-text conversions emphasize precision and accuracy. And customers are responsible for driving accuracy.
- Unsupervised or semi-supervised ML model training- Only several hours of data is captured, enough for fine-tuning the ML models.
- Cloud computing- By using Cloud computing, data processing speed is boosted without any lags in processing.
