Hi everyone, my name is, thanks for joining to my session today on Microsoft Cloud Agent Insurance And Greetings. Coming to my session today. Today we are going to discuss about how the Microsoft Cloud AI NextGen Insurance, and we'll be discussing about the core topics like the core machine learning morals and discuss assessment under neural networks in insurance, natural language processing, in document analysis, e and LP, performance metrics and predictive analysis and fraud detections. And computer vision, performance metrics and integrated a platforms and automation, economic impact of a insurance and future of yay and insurance underwriting, SharePoint, AI transformation collaboration. And these are all the topics that we are going to. Talk in this session and how we see between this traditional methods and as well as this transformation between this, what are the benefits, what is the accuracy and what is how the under underwritings are going without the paperless, we'll all see this features, session this Microsoft Cloud AI next generation session. The insurance industry is experiencing performed digital transformation. Through a integration fundamentally reshaping traditional business models and operational processes with the global insurance, yay Investments, and it's reaching 21 billion in 2023, and the predictions of 45 billion in 2027. The insurance industry is recognizing a potential to revolutionizing the core processes in today's world, the insurance carriers implementing AI driven underwriting systems. Having a reported of 40% risk reduction in the risk assessment time, and 60% of decrease in the processing cost and 35% improvement in the loss ratio production. Sir, these systems can these systems can also analyze over 200 unique data points per application in real time. Compared to 40 to 50 points in the traditional underwriting, and this is all about the key highlight of the insurance in the industry, then now we will go by one section how we can leverage the benefits of using this gen features in the insurance coming to this core core machine learning models. In the risk assessment. Here we talk about, three models. One is a random forest, one is a logistic regression, and one is the gradient boosting mechanic. When coming to this random forest the achieving the accuracy rate is like above 85% in risk classification. Significantly outperforming that traditional actuarial methods can process up to 75 unique features per assessment with the optimal performance through hundred to. One 20 decision enables and coming to this logistic logistic regression, this logistic ion maintains over 78% of accuracy rate while providing the complete algorithm transparency for regulatory compliance, delivering 28% in the improvement of the risk. Of the risk of segmentation accuracy com compared to the conventional methods when it come to the gradient boosting mechanism, boost things implementation. This kind of machine learning model achieve 90% of perception in the risk classification. While maintaining the real processing speeds can efficiently analyze over 300 distinct features per application In. 30 seconds. Now coming to the next topic of the neural networks in insurance, the neural networks particularly is nothing but an advanced learning architectures have transformed how the insurance processes of the unstructured data during the underwriting, this cutting edging conventional na natural networks deployed in properly incidents Now identify. Potential risk factors from the visual data with above 91% accuracy, seamlessly processing 77 50 property images daily, while dramatically reducing the need for the human review, the deep learning experience, the deep learning excellent unstructured data processing and structured data analyst. We'll see how it happens, the accuracy rate and the processing and the risk classification. In each here in the deep learning excellence, we will see over 91% of accuracy in the risk assessment, unstructured data processing seven 50 images daily. This is what, how the neural networks, if we implement certain in the insurance practices, we can achieve these targets. Like a structured data analysis. Over 86% of risk class classification accuracy. This is how the neural networks that helps in the insurance industry for the deep learning, unstructured data processing, unstructured data analysis with the accuracy rate. Now coming to the natural language processing in document analysis, this is very important because the reduction in reduction in manual review time. Be model accuracy named entity rec recognition accuracy. So now coming to this reduction in manual review term, it's not only about how the AA features are utilizing it also should have a manual review review term reduction, and that how source this in LP system process, like a tool k plus. Pages daily, over 300 pages when we compare with the traditional manual methods when coming to this bet model accuracy for extracting the critical information from the policy documents and when it comes to the na named entity recognition. Accuracy. It is about 94 percentage in identifying and categorizing the key entities with the insurance documentation. The insurance underwriting's reliance on document processing has been revolutionized by the NLP models. Be based models, trained on 2.5 million insurance. Documents can analyze a standard 25 page policy in just a few seconds, just like 20 seconds comp compared to 52 minutes of the traditional methods while maintaining accuracy. While maintaining the accuracy levels. About 85%. Even with the complex riders and endorsements, and this is one of the beautiful natural language processing in the document analysis that helps in the insurance industries though, N-L-P-N-L-P performance metrics in. Insurance. The here, the main five topics that we models we are going to talk is like the beta document analysis, technical inter interpretations, the cover gap detection and the NER systems in coming to this section. The first topic, like B-R-T-B-E-R-T document analysis, over 90% of accuracy rate with 20 seconds around like eight 18. 18 to 20 seconds per document and the 85% reduction in compared to the manual processing. The technical term interpretation. 93 per 93 percentage of accuracy rate processing over eight 50 documents per 72% of reduction in the processing time. Cover gap detections, the accuracy rate will be over 90% and eight 50 documents processed to cost to savings will be like a three, two. $4 per document. Iner systems accuracy date will be 94% rate and 8,000 documents processed daily using this INER system. And the cost savings also like amazing, like a four around like five $5 per document. Pre coming to the next section, how this pre predictive predictive analysis and the fraud detection this in this things, in this predictive analysis on fraud detection, we will talk about like advanced predictive models time series analysis and graph neural networks. And auto auto encoder integrations. This this advanced advanced predictive models. We see this reduction. It is just like 36% reduction in the combined loss ratios and 41 point two percentage improvement in the risk assessment precisions the time series analysis, over 84.8 percentage of accuracy in predicting claim. Occurrence within the first 24 first 12 months of policy insurance graph neural networks, over 71% of implementing fraud detection rates while reducing the false positives by 62%. Auto encoder integration. Over 90% of accuracy in identifying unknown mollusk claim patents while processing 65,000 claims daily. And when we come to this predictive analysis on fraud detection, this plays a key role. This plays a key role when we are using the a AI features in the insurance ancy or in any other industry that deals with the most of the content management and the document management where the A underwriters or the officers will be using on today, day-to-day basis for collecting the meta content and metadata from the actual business users. Coming to the next topic, the computer Vision in property assessment in, and this is not only about the insurance. We have something like property casualty, specialty casualty, the Iron Show in a different models of the insurance. When it comes to the property assessment, the computer with the AI feature, the computer vision in property assessment has like drastically showing the improvements. One is the image capture. One is the cost estimation, one is the future detection and one is the damage assessment. It's not only about a small property like a home or this it also deals with this like the, a deals with the the big system like the very San Francisco Bridge or the New York Bridges. The, were big malls in the New York, the big buildings in the New York and San Francisco and other parts of the country. When we have a huge like billions of the properties. And that also helps this AI also helps in that assessing with through the computer vision in the property. Assessment. One is image captured here in this image capture. We'll discuss it process inserts with the ultra high resolution property images enabling the detailed analysis. They're now coming to the feature feature detection. Advanced yolo VFI and Foster CNN algorithms identify 180 5 distant property features with 87%, 87.4% perceptions and damage assessment. Specialized CNN models achieve over 90% accuracy in identifying structural damage, CBRT and extent. Revolutionary, revolutionary computer vision systems have slashed the property assessment time by 70%, over 70%, while simultaneously initially boosting the damage detection accuracy by 41%. This high performance system processes can impress you like around nearly 2000 of property measures daily. This deliver delivering comprehensive assessment results in me 4.2 minutes per. Property diametrically faster than 52 minutes required for the traditional human inspections. The computer vision performance metrics, it's not only about computer vision in the management system, it be how to check, like how the computer vision has the performance metrics, how we can calculate that, how we can, because the metrics is the one thing that shows the graphical representation where to compare with the previous data and the future data to predict that metrics. This coming to this section, this is like a CN instructional damage. Assessment, security system, fire assessment fire risk assessment. Yay. Virtual traditional inspection over 90 nearly 90% of accuracy rate in identifying structural damage, severity and extent, and comprehensive assessment in just eight point 8.5 minutes. Per property security system detection industry leading over 90% of accuracy rate and 52% of implement in future detection compared to manual methods. The fire risk assessment, over 90% of accuracy in identifying potential fire hazards. Part of 85% reduction in the assessment. Yay. A virtual a versus traditional inspection. And this computer vision is having a accuracy rate for around 90% and manually towards 70 or above 70% accuracy rate. The average costing can, it can be around between 200 to 300 range. The computer, advanced computer vision algorithms have revolutionized the property evaluations. Slashing assessment times by 84%, went from 50 minutes down to the, just the 8.5 minutes or nine minutes per property, while maintaining consistently high accuracy rates across all the assessment categories, integrated AI platforms and automation in this integrated platforms. And automation. The robotic process of automation one is policy renewal processing the price optimization market segment optimizations. Here in this robotic process of automation, it achieves industry leading 97% is accuracy in automated data entry while processing of about 7,000 documents daily. Slashing manual processing time by 72 point. 5% and maintaining the error rates below 0.5%. And when it comes to policy renewal processing, it's not about only the policy or the claims or the underwritings it's also about the renewal. Process, right? Any policy is getting renewed within a 12 months or 18 months based on the policy inspection date and the policy inspection end date. So the policy should get automatically go for the renewal, so the dramatically cuts processing times from 22.8 hours to just 18 minutes per policy while sustain sustaining 98% accuracy, efficiently handling. Close to 9,500 complex renewals monthly market segment optimization provides exceptionally over 88 percentage of predict to accuracy for customer responses to price adjustments across diverse market segments, directly contributing to 2.3 percentage point market share growth. We'll also see about we, we have seen about the computer vision, how the platforms and how the, a integrated platforms and how how the neural networks helps. Now, we'll come like it's not about always. Leveraging the AA features or any other technology. It is also important to see how economically IT impact of AA insurance when it comes to this operational cost reduction, processing efficiency, and customer experience. This is the most important when it's not only about. Implementing a features or gene features a new AA feature, but it's also important to see how how economic impact of a insurance, the operational cost reductions, integrated AA platforms resulted in average operational cost reduction of three point. 8 billion annually with the return of investment. Typically assured within 1818 months of the deployment processing efficiency, if you see the processing efficiency, the reduction of around reduction of 58% in end to end processing time for the standard insurance operations while improving the overall accuracy by 41% and processing of 12 k plus insurance transactions daily. When it comes to the customer experience, at the end at the end customer experience. Is the most important thing, and here we'll see how by utilizing this economic impact of AI in, in insurance. So the carriers utilizing this, the carriers utilizing this AI systems has reduced policy processing cost by 62% while improving the customer satisfaction metrics by 35% through faster response time and reduced error rates. The future of EAA, the future of EA. Insurance underwriting, current implementation near term elu evolution. The future transformation. The current implementation is like the operational efficiency, and the cost reductions near term evolution, dynamic risk assessment and personalizing. Pricing the future transformation. Fully automated underwriting. Fully automated underwriting with a continuous risk monitoring. The integrated the integration of the integration of artificial intelligence in insurance underwriting represents a paradigm shift in the risk assessment. Under the policy management methodologies. This transformation extends beyond me automation fundamentally. Altering for altering how business insurance companies evaluate risk process, document the fraud detection, and interact with the customers. All these technologies continues to evolve. The insurance industry is positioned for further innovation, particularly in areas such as personal personalized pricing. Dy, I'm sorry. Dynamic risk assessment and automated underwriting. The future will be, the future will likely see even greater integration of AI technologies leading to more sophisticated and efficient way of operations. That better both carriers and policy holders. And coming to the topic like the SharePoint yay. Transforming collaboration. If you talk with insurance, we should talk about SharePoint. That's how it is tightly integrated. The technology integrated with insurance, because insurance is all about the content Metadata. And the file repository, the SharePoint here, that the day, the SharePoint bond, it has has the come up with the features with each version, and it's trying to help the insurance, the life sensors and the construction in industries the most of the tough way in. Industries when it comes to this tofa industry, the always the insurance industries will place the first with the SharePoint, where it has the best, the Microsoft SharePoint provide the best repositories, the cloud repository, the collaboration, and the collaborate and team collaboration. Platform for leveraging this share pan features like the automation Yay, power. Now, the now coming to that the latest feature of this share pant, yay. How they transforming the collaborations. And in today's world, like the Microsoft yay, a integration is revolutionizing again, the share pant capabilities. It's not only about the SharePoint, the Microsoft itself is coming up with this, all the products by the AI integration. Okay. The delivering measurable, productive gains and transforming the collaborative workflows through intelligent features that adopt to our organization needs. Co-pilot in co-pilot in SharePoint co. The introduction of this the co-pilot or integration of this co-pilot in SharePoint provides contextual assistance. For the content cre, for the content creation knowledge discovery, the real time data summarization within SharePoint, reducing this information ritual by 40% while enhancing the decision quality through personalized insights and recommendation. Now come when we talk about the AI power authoring, the enhance, it enhances the document it enhances the document creation. With the intelligent suggestions, right? That's how the authoring will happen, right? EA powered authoring, it enhances the document creation with the intelligent suggestions like a contextual formatting and tailor the content recommendations that align with your organization, communication patterns and the institution knowledge, increasing the content quality and the consistency up to 35%. Coming to the intelligent content management. So the, it automates the content classification, which improves the searchability and delivers the personalized content experiences based on the user behavior and the organizational context, resulting in 28% efficient information. Discovery under knowledge sharing. Now we'll talk about AI powered intelligence agents and automation. So these a powered intelligence, the agents and automation, how it plays a key role in the insurance domain. So the coming to the SharePoint agents like the intelligence agent. Autonomously. Handle autonomously. Handle autonomously, like it's independently, handle the routine task. Okay. While continuously learning from organizational patterns, these AI entities proactively monitor content anticipate the needs and deliver the actionable insight. Based on the collective knowledge reducing administrative overhead by up to 45%. The SA workflow automation, this AI driven process orchestration identifies the bottlenecks and the and implements optimized workflows in the real time. This one, automation eliminates up to 60% of the manual interventions, ensuring the consistent quality and adopts for changing business requirements without restricting this operations. And thank you for joining my session, and thank you for joining my sessions and I'm having like around 10 plus years of experience working with this. It industry, particularly with the insurance domain. If you have any questions, if you have any, if you need any PS or related like how you want to move the content from the source to the cloud and how you want to leverage this AI features, how you want to plan your strategy with the SharePoint to moving the content, the versioning, the collaborations between this insurance remains. And the technology, you can always reach out to me. You can. I'm very happy to help you and thanks again for joining my session and see you. Thank you.