Puravankara Dials Up AI to Boost Call Centre Performance and Customer Engagement

Founded in 1975, Puravankara Limited is a real estate company headquartered in Bengaluru. It also has a presence in other Indian cities, including Chennai, Hyderabad, Pune, Mumbai, Kochi, Goa, Coimbatore, and Mangaluru.

The 50-year-old company operates a 100-seat call centre, serving as potential customers’ initial touchpoint. The facility assists, resolves issues and handles inquiries, thereby playing a vital role in shaping the company’s overall customer experience. With such a crucial business success parameter tied to it, Puravankara decided to streamline its operations.

“I strongly believe that without a good employee experience, One can’t achieve the best customer experience, EX and CX go hand-in-hand. So, with an eye on ramping up both, we decided to reduce the human effort and investment required for call interaction analysis, remove biases in the call analysis process, bring transparency and increase the accuracy of the call analysis”, says Arvind Singh, Chief Technology and Product Officer (CTO) at Puravankara Group.

To achieve these key objectives, drive greater efficiency and effectiveness in the call analysis workflows, Singh opted to harness the power of AI.

Implementing AI for Voice Analysis

Singh brought an AI solution startup Convin.ai for call analysis to inspect and understand the customer interactions, it was rolled out in November 2023. He used the vendor’s AI model and then fine-tuned it based on Puravankara’s business needs.

Fine-tuning existing AI models with internal datasets and business nuances is an essential step in improving the quality of any AI solution. While fine-tuning AI models is often seen as a daunting task by ICT groups while just starting on their AI journey, IBRS notes that significant improvements have been made in the ease and speed of performing such activities in the last six months.

“Implementing an effective AI solution requires extensive training, tweeking and re-training to align it with specific business needs, as every organisation has a different business model and strategy, and is at a different growth phase. We trained the AI on our dataset and are still continuously refining the AI model to reach a higher level of performance and accuracy. Though it’s unlikely to be 100 per cent accurate, even if the AI can achieve 80 per cent effectiveness, it can still be of tremendous help, by flagging the anomalies and interactions that warrant further human analysis and review”, Singh says.

The company has further enriched its overall dataset quality and sufficiency through its sales team. After each face-to-face interaction with a potential customer, the sales representative captures and feeds the details of the conversation into the system, which helps to improve the AI’s capabilities over time and creates a virtuous cycle of improvement.

“Through the AI solution we implemented a year ago, we are analysing all customer interaction recordings to identify patterns and insights. The AI algorithms help us pinpoint specific behaviours, quality of interaction, information flow and transparency, keywords oftenly getting used that correlate with positive or negative outcomes. This allows us to understand factors like how often customers reference competitors, repeat certain words, or exhibit other behaviours”, says Singh.

Spike in Customer Delight and Employee Performance

By leveraging the AI-powered analysis, Puravankara is continually refining and optimising its customer service approach. The initiative has helped the company in multiple ways.

In the traditional world, call centres go through all the audio files manually. They hear them, audit them, and then come up with findings. However, it is nearly impossible for a large and old organisation like Puravankara to review all audio files accumulated over the years.

“Such a manual process is also plagued with bias. Those auditing might choose to leave out some people, but this platform removes that bias. It will audit all the recordings against the defined algorithm and deliver an unbiased result that will guide you to know what is happening on the ground, without any biases”, Singh says.

A potential lead who represents a future revenue opportunity for the business may have multiple questions that must be addressed effectively. He/she should understand the brand, receive the right information, and have their questions handled smoothly through an interface that enables the customer service team to comprehend what the customer is seeking and who they are.

With AI providing the background context to every customer and additional relevant information, it empowers the call centre executive. It prevents the customer from having to repeat the same information multiple times, which can lead to frustration. The call centre executive is also able to pick up the discussion from where it was left off, rather than starting from scratch.

IBRS recommends organisations running call centres review available digital voice AI services. They offer easier self-service capabilities, allowing customers to find the right solutions for their concerns instead of waiting for a live agent. This also provides them with a sense of autonomy and control over their interactions with the brand. AI-powered voice agents are trained to understand specific intents that customers may have when they call. They are designed to recognise even subtle nuances of human speech, enabling them to provide quick and seamless assistance without requiring human intervention. However, enterprises need to invest in purpose-built, voice-based touchpoint platforms that can handle human voice conversations effectively.

The third critical aspect for Puravankara involves identifying employees’ training needs. Analysing performance — why some excel and others struggle to convert leads to site visit – and site visit to booking. This allowed them to identify training needs for those who need improvement and incentivise those who perform better. Sharing relevant snippets from these discussions can guide us. When the system points out areas for growth, individuals often accept it more readily than when a human delivers the same feedback, avoiding emotional biases.

“By leveraging AI to enhance our customer interactions, we have seen a 10 per cent improvement in conversion rates as the quality of our conversations has increased. Customer delight metrics have risen by an estimated 20 per cent, while individual employee performance has gone up by 25 per cent”, Singh says.

IBRS sees Singh’s work with AI as a prime example of what is possible and highly effective. To leverage the power of AI ethically, IBRS recommends that enterprises should set up their own AI development policies and checklists as an essential act of governance. Management must educate itself and all corners of its organisation (including human resources) on AI issues, including customer acceptance and safety along with the critical business risks of AI in their organisation.

Future Plans

Singh wants to next leverage AI enabled CopiIot and co-assists in several other areas.

He aims to introduce Copilot for the leadership to get real time business insights, co-assist for the sales team, where it can provide real-time answers to questions and serve as an interface for immediate discussions. The second use case Singh has in mind involves leveraging AI for the hiring process — from filtering resumes based on capabilities to identifying candidates during group discussions to onboarding them. 

Not only this, “AI can also assist by efficiently analysing contracts, flagging relevant sections with anomalies, and improving accuracy. We also intend to use AI in cyber security as it can step in by identifying anomalies, suggesting solutions, and making the tools more self-learning and self-healing”, Singh adds.