Introduction

• In this episode of our podcast, we had the pleasure of hosting Pranshu, an alumnus of IIT and IIM, and a seasoned expert in data science, insurance, and business transformation. With nearly two decades of experience and leadership roles at Capital One, Aditya Birla Health Insurance, Ola Insure, Paytm, and several startups, Pranshu has worked at the intersection of data, digital innovation, and insurance reform. In this blog, we unpack his powerful insights on what’s broken in traditional insurance, where AI can deliver transformational value, and how to reimagine insurance from the ground up using intelligent systems.

• Listen Full Episode :- (https://youtu.be/LV2QShR4BuQ?si=7HMikMg7SLT3zsra)

What's Broken in the Insurance Industry?

• Pranshu begins by comparing the innovation curve across financial services, highlighting that while payments, lending, and wealth have embraced tech transformation, insurance still lags behind. “Insurance,” he says, “is the poor cousin in the family of financial services.”

• Core challenges include:

• High cost to serve: With 60–70% of payouts going into claim ratios, operational and distribution costs create additional inefficiencies.

• Static products: Insurance offerings are often designed without contextual relevance to evolving customer needs.

• Limited personalization: Pricing and underwriting remain rigid, based on outdated assumptions.

• Sales over science: Instead of focusing on actuarial strength and data models, insurance has become overly focused on distribution channels.

• These inefficiencies not only affect profitability but also hinder customer trust and adoption.

How AI Can Revolutionize the Insurance Value Chain

• AI has the power to create value across the insurance journey from acquisition to claims. Pranshu outlines a vision where AI doesn't just automate tasks but transforms how insurers think:

Key Areas of Impact:

• Customer acquisition: AI enhances targeting and engagement strategies by analyzing behavioral, demographic, and contextual data. This allows insurers to identify high-intent prospects, predict timing for purchase decisions, and optimize outreach channels. Instead of using blanket campaigns, companies can now deploy tailored messaging based on the user's journey, increasing lead conversion and lowering marketing spend.

• Personalized pricing: Traditional insurance pricing is rigid and often based on broad actuarial assumptions. AI enables insurers to offer personalized premiums by leveraging real-time data such as driving behavior, fitness activity, lifestyle habits, and spending patterns. This leads to more accurate risk profiling and dynamic pricing models, allowing customers to receive plans that reflect their actual risk levels and preferences.

• Fraud detection: AI excels at identifying patterns and anomalies in large data sets, making it an effective tool for fraud detection. By analyzing historical claims data and comparing it with new inputs, AI can flag suspicious claims in real time, assess authenticity, and recommend escalations. This proactive monitoring helps reduce false claims, protect the insurer's bottom line, and streamline investigation workflows.

• Claims management: Traditionally, claims processing is a manual, paper-heavy task. AI changes this by automating document processing, verifying claims data, and assessing eligibility using trained models. Chatbots and voice agents can guide users through the claim submission process, while backend algorithms validate, approve, or escalate cases instantly reducing turnaround times and improving customer satisfaction.

• Retention: AI-powered predictive analytics can identify customers at risk of churn by analyzing engagement patterns, satisfaction scores, claims history, and life events. Insurers can then deploy personalized retention strategies such as policy adjustments, timely offers, or improved service touchpoints to maintain customer loyalty. This also enables a more proactive approach to customer relationship management.

• While all functions stand to benefit, Pranshu highlights distribution and claims as the two most influential levers:

• Distribution: AI increases conversion while reducing CAC.

• Claims: Streamlined processes directly lower the insurer’s cost burden.

Personalizing Distribution with Generative AI

• Insurance distribution is undergoing a dramatic shift from transactional selling to intelligent conversations. Pranshu emphasizes how generative AI now allows insurers to understand context, behavior, and customer emotion in real time.

• Examples he cited include:

• Journey-aware conversations: If a customer is at the airport, AI can predict travel anxieties and tailor travel insurance offers accordingly.

• Conversational bundling: Systems can now suggest plan combinations dynamically based on behavioral signals and past purchase data.

• AI storytelling: Generative AI crafts micro-targeted narratives that resonate emotionally, improving engagement and conversions.

• Counter offers & dynamic pricing: Real-time decision engines tailor prices based on affinity, risk profile, and even purchase momentum.

• AI turns distribution into a personalized, continuous experience not a one-time sales pitch. According to Pranshu, the real innovation lies not just in having the tools, but in applying them meaningfully.

Evolution of AI in Insurance: From Models to Intelligent Agents

• Pranshu reflects on the evolution of data science over the last 10 - 15 years:

• In 2008 - 2013, insurance companies hired statisticians to build complex risk models manually.

• By 2016 - 2018, the emergence of pre-trained machine learning models and faster computation made it easier to deploy solutions.

• Today, generative models and intelligent agents allow near real-time, no-code model creation and automation across the stack.

• Tasks like digitizing unstructured data, rule-engine building, underwriting, and claims can now be performed autonomously or semi-autonomously. AI systems can adapt dynamically to feedback, iterating faster than human teams and enabling granular segmentation that was once impossible.

• This leap in capability reduces the need for large data science teams and makes experimentation affordable and fast.

Understanding ROI from AI in Insurance

• When asked about the ROI of AI, Pranshu drew an insightful parallel to the early internet era:

• "People questioned the ROI of the internet. But once we had layers like e-commerce, video, and mobility, the ROI wasn’t a question it was obvious."

• AI is in the same phase today. The foundation models exist. Tools are usable. Infrastructure is mature. What’s missing is the application mindset.

• To extract ROI from AI:

• Set AI KPIs at the leadership level

• Align tech and business around real customer pain points

• Encourage a culture of experimentation over rigid playbooks

• Measure success by customer impact, not just automation efficiency

What Would an AI-First Insurance Company Look Like?

• If Pranshu were to start an insurance company today, it would be radically different from incumbents. It would be lean, intelligent, and built on API-first thinking. Here's how he would approach it:

1. Demand Intelligence

• Focus on understanding anxieties, not product categories

• Replace form-filling with conversational bots that understand intent

• Personalize offerings based on lifestyle, moment, and context

2. Smart Supply

• Build seamless connections with hospitals, garages, and care networks

• Enable instant pre-authorization, policy verification, and approvals

• Replace documentation with smart APIs and AI-led decisioning

3. Dynamic Risk & Pricing Engine

• Use real-time risk signals and customer sentiment to price policies

• Adjust prices based on usage, behavior, and willingness to pay (like Ola surge pricing)

• Constantly iterate pricing models at micro-segment levels

• Such a company could be run by a small team and serve millions of customers without massive operations or legacy tech baggage.

Can AI Improve Insurance Penetration in India?

• India’s insurance penetration remains below 3%, and AI can be the key to unlocking the next growth wave. As Pranshu noted, the issue is not the demand its distribution and servicing cost.

• AI can reduce the cost of serving low-premium customers, making it viable to target:

• SMEs and micro-businesses with parametric and embedded products

• Rural and Tier 3 markets with voice-based agents and local vernacular tools

• Digital micro-insurance plans offered contextually at point-of-sale (e.g., a wine pack, shop purchase)

• Bite-sized insurance plans, enabled by AI, make insurance accessible, flexible, and scalable for India’s underserved segments.

How Should Insurers Adopt AI?

• Pranshu offered a practical and proven roadmap to help insurers move beyond AI experimentation and into real-world adoption:

• 1. Create AI accountability at the top: Successful AI transformation starts at the leadership level. Pranshu emphasizes embedding AI goals directly into the KPIs of CXOs and business heads. This ensures strategic commitment and visibility. When AI becomes a boardroom agenda item, it receives the budget, prioritization, and cross-functional backing it needs to scale.
• 2. Encourage cross-functional collaboration: AI doesn’t live in isolation. It requires collaboration between business, technology, product, and operations teams. Pranshu suggests breaking traditional silos and creating AI pods where diverse team members work together, often informally, to identify high-impact areas. AI success stories often start over whiteboard sketches and hallway conversations.
• 3. Start from the customer backwards: Rather than digitizing internal processes blindly, Pranshu encourages teams to think from the "moment of truth" in a customer’s journey. Whether it’s buying health insurance for a parent or filing a claim after a car accident, start with the pain point and work backward to see how AI can make the experience faster, simpler, or more human.
• 4. Launch fast, iterate faster: In the age of generative AI, building proof-of-concepts no longer takes months. Insurers must move from long requirement documents and rigid PRDs to quick prototypes, MVPs, and real-time user testing. With low-code and no-code tools, frontline teams can experiment directly bringing ideas to life within days instead of quarters.
• 5. Measure success in experience, not just automation: It’s easy to fall into the trap of measuring AI success by back-office efficiency. But Pranshu argues that the true metric should be customer delight. Did the AI reduce anxiety? Did it make the buying or claims experience smoother? Insurers must shift from operational metrics to experience-driven KPIs.

• This approach not only helps organizations adopt AI faster, but also aligns the technology with the core mission of insurance, protecting people in moments that matter. Pranshu offered a practical roadmap for AI adoption:

1. Create AI accountability at the top: Make it part of your leadership KPIs.

2. Encourage cross-functional collaboration: Bring business, tech, and product teams together.

3. Start from the customer backward: Build from real-world moments of truth.

4. Launch fast, iterate faster: Replace PRDs with prototypes and experiments.

5. Measure success in experience, not automation: AI should enhance how customers feel, not just reduce ops cost.

• He emphasized that the biggest bottleneck isn’t technology, it’s imagination and organizational culture.

Is the AI Wave Real or Another Hype Cycle?

• Pranshu has seen neural networks, random forests, and big data cycles come and go. But this time, he believes, AI is finally in the hands of everyone:

• Product managers are building UIs using Bold.ai and Lovable.

• Kids are asking ChatGPT to write personalized bedtime stories.

• AI is being used to design, iterate, and launch in hours not weeks.

• This democratization means we’re not just crossing the chasm, we’re in the middle of an adoption boom. The only question is who will move fast enough to take the lead?

Final Takeaways from the Conversation

• Here are Pranshu’s top lessons for the future of AI in insurance:

• AI augments humans, it doesn’t replace them

• Claims, distribution, and underwriting are the biggest levers for impact

• Small teams with smart agents can build billion-dollar platforms

• AI should serve real customer moments, not just backend tasks

• Success lies in experimentation, not perfection

Want To Listen Full Episode ?

• This blog only scratches the surface of Pranshu’s insights.

• Listen Full Episode :- (https://youtu.be/LV2QShR4BuQ?si=7HMikMg7SLT3zsra)