Transforming into an AI-First organization
Mehmood Mansoori, President - Shared Services & Online Business at HDFC ERGO delves on leveraging AI and reaping benefits:
Ravi Lalwani: Briefly describe 3 important AI projects initiated during the lockdown and what were the business objectives of these projects?
Mehmood Mansoori: One of the things that covid has taught us over the past year is the power of the internet. With the country in complete lockdown during the initial days of the pandemic, digital adoption became the way of life. HDFC ERGO has been at the forefront of driving innovation in AI for its customers and the pandemic gave us further impetus to enable our customers by putting the power of digital and AI in their hands. Some of the important initiatives we took forward during the lockdown are:
Servicing customers through AI-enabled bots: We offered several services to our customers through digital platforms to ensure seamless servicing, as well as to ensure their well-being while they stayed at home. These services were offered through AI-enabled bots like DIA (Digital Insurance Assistant), WhatsApp bot, and eMail bot (eRA), which ensured real-time servicing 24x7. Today, more than 55% of customers’ requests are serviced digitally through these platforms and in the last 6 months, despite all constraints, we have witnessed an upward trend in our NPS.
AI-based motor claims processing:
On the motor insurance front, we launched the AI module IDEAS (Intelligent Damage Detection Estimation and Assessment Solution) for the settlement of claims. The IDEAS module uses neural network image processing & analytics, machine learning, and natural language that supports instant damage detection, as well as calculation of the claims and estimate for the surveyors to help in motor claims settlement in real-time.
WhatsApp & Telegram: To take our services to tier 2 & 3 towns where internet penetration is good, but consumers are not so comfortable using mobile apps or selfservice portals, we launched our natural language processing-based services on WhatsApp and Telegram, which turned out to be a big success.
Which objectives were easily achieved, and which were tougher?
Our key objective was to offer a hasslefree experience to our customers, be it at the time of buying a policy, renewing it, or at the time of claim. We are proud to say that we successfully managed to achieve it.
Given the complexity of requests involved in the insurance domain, it is a daunting task to use bots. For instance, when a customer writes us an email for a service, there is a complexity involved and it is not as simple as requesting a cheque book or a bank statement. Moreover, our every response is a legal commitment to our customers, which means our solutions need to be accurate. Further, when the customer approaches us, especially during the claim, it becomes imperative to demonstrate empathy. We have trained our bots with the help of behavioural scientists, which have also matured our AI solutions.
We are overly excited about the possibilities and investing the best of our talent in such initiatives. Now, we have several high-impact use cases. Around 30% of our digital servicing is catered through AI, which is making a huge difference. Our turnaround time has improved significantly, which is adding to our operational efficiencies.
What technologies and technology companies were associated with the projects?
We have a robust IT infrastructure and team and believe in building most of the solutions in-house. But we are always on the lookout for new technologies and hence have partnered with some leading tech companies and start-ups across various domains. We are investing in the right talent and training internal resources with the right aptitude in AI technology and moving to become an AI-First Company.
What are the main internal and external data sources for the projects?
With a customer base of more than 15 million customers, we have sufficient data for AI. We also use previous data of customer services at various touchpoints to train our AI models.
Briefly describe the team structure, size and members that developed, tested and rolled out the project.
Backed with an extensive team of professionals, we have invested hugely in technology. Also, we have a dedicated digital innovation lab where around 30 sharp minds are continuously exploring emerging technologies and experimenting on various use cases. Once the lab finds the right fit, it is handed over to our technology team who runs a project and takes it to the production stage.
Briefly describe the challenges faced during the project and the lessons learned.
To work on AI technologies requires patience. One must experiment with the vast amount of data intelligently and train models. Further, the resultant technology may also need to be fine-tuned with business realities. Also, for an effective AI, one may need to tweak business processes as well. All these require not only sharp tech minds but creative business champions who can think of out-of-the-box, use cases, and work closely with tech teams to help implement solutions.