Exclusive Interview with Satyakam Mohanty, Head of Leni and Mosaic Applied Intelligence at LTI

AI-based decisions enable people to rethink how to integrate information, analyze data, and use the resulting insights to enhance decision-making. Artificial intelligence-based solutions form the basis for all computer learning and are the future of all complex decision-making. In the era of digital transformation, AI-based decision-making has become important for every business to compete in the market and to fast-track its operations.

Speaking with Analytics Insight, Satyakam Mohanty the Head of Leni, and Mosaic Applied Intelligence at LTI sheds light on Leni, an AI-powered decision-making platform, and its augmented insights.

What is LENI?

Leni is an AI-powered decision-making platform built for business users with the vision of "democratizing" decision support at scale. Leni's approach to augmented insights focuses on the business decision to be made and works backward to connect the user context with relevant analytics and autonomously generated data narratives by uncovering deeper, more valuable insights with ease.

By using Leni's AI and machine learning capabilities, business users can apply advanced analytics interventions like key driver analysis, predictive analytics techniques like forecasting, and prescriptive analysis like simulations to their data in a simple Q&A mode. Being a data format platform and domain agnostic, Leni can ingest, analyze, interpret, and explain data and provide lightning-fast insights to decision-makers.

Why was it developed?

Many businesses struggle to unlock value from their data and analytics initiatives because of the overwhelming possibilities of analysis and surging data volumes. The existing BI and analytics marketplace serve two very distinct needs of business users which can broadly be classified as:

Everything you need to know: All the BI, MIS & dashboards that business users can access to understand their current performance.

Everything you need to do: All the strategic projects that are being worked on by the data science & analytics team to drive business goals.

What is missing is the ability for them to get the answers to the most critical questions and make informed decisions like why something happened, what it means for the business and what's next, etc. Such insights are not readily available because they can only be achieved with the help of specialist data science teams generating multiple reports and dashboards. These are very difficult to comprehend by a non-data-savvy business user.

Leni is designed to solve this gap keeping the user context and insights comprehensiveness at the center.

What are its main characteristics (including the technology and how it works), and why do they matter?

Leni's augmented data discovery, autonomous insights, and intuitive interface is designed to solve business user's need for speed, accuracy, and comprehension. Leni's deep learning-based insights recommendations give the freedom to business users to chart their insights journey, whereas the autonomous nudges provide the 'peripheral vision to business users to spot changes, trends, and anomalies. From data preparation, connecting to datasets to performing and interpreting the analysis itself, Leni does all the heavy lifting and provides the most accurate and relevant insights to the business users.

Built with custom and patented technology, Leni's architecture is built from the ground up to provide speed, scale, and innovation along with enterprise security.

Can you take us to through how it works?

Smart Search- Smart search powered by NLP empowers business users to ask questions in plain English and get accurate responses with rich narratives and visualizations in real-time.

Full Spectrum Advanced Analytics – Full spectrum of analytical complexity is accessible through NLP search, at the click of a button, truly democratizes analytics for end decision maker

Simplified Consumption – Autonomous nudges detect deviations that are likely to result in potential disruptions as well as identify reasons for deviation and NLG powered narratives with rich visualization and explanations (Explainable AI) provides context and directions to insights

Flexible Setup and Deployment – Leni's creator module can ingest data, configure metadata & ontologies, set up KPIs and advanced analytics models – all in less than an hour for businesses to consume. Once a solution is configured on Leni, the platform training enables federated learning across use cases and industry verticals, with no additional time or effort.

When was it launched?

Leni was launched in 2017, as the flagship product of Lymbyc, a specialist artificial intelligence, machine learning, and advanced analytics company. The company was later acquired by LTI in 2019. LTI and Lymbyc were strategic partners for over a year before that, having collaborated to deliver advanced analytic solutions to global enterprises.

What advantages does it accord to insurers, including metrics about improvements they can expect?

Leni's AI-powered autonomous intelligence helps underwriters, brokers, and insurers mine, measure, manage and monitor critical KPIs', improve underwriting performance, and broker profitability with self-service analytics. Using Leni, insurers can directly pose their queries to Leni for getting real-time insights with simplified narratives to help them take effective business decisions- thereby decreasing dependence on analytics teams and reduced time to insights.

Advantages to the insurers:

• Understand levers of profitability and expenses and identify root causes with deep-dive diagnosis.
• Connect to disparate external sources such as Open source (HIFLD, FBI), Hazard Hub, Pitney Bowes, and intuitively search prospects to generate leads for underwriting new business.
• Identify hotspots in their book of business in terms of quote conversion and last 12-month loss ratio without having to spend additional time to build it.
• Simulate optimal retention and loss ratio using what-if feature and track key anomalies in loss ratio, renewal, and GWP trends across multiple dimensions.
• Gauge the impact of KPIs such as claims severity, the premium earned on target profitability and foresee the likely number of submissions or bind rates in the future for optimal resource allocation.
• Track and score the performance of brokers based on multiple parameters such as commission % earned v/s revenue generated, average policy size, strike rate to keep a close watch on broker productivity.

Expected Improvement

• Query resolution time reduced from weeks to minutes – Reduce dependence on data science teams for advanced analytics interventions.
• 90% accuracy in model analysis.
• Freedom of exploration with a self-service advanced analytics platform.
• Redefined user experience with natural language search interface.
• Empower cross-functional teams with real-time risk, profitability, and broker-related insights.

Can you please provide an illustrative example of how LENI works?

A leading P&C insurer's business group was grappling with a longer turnaround time for resolving any analytics-related ad-hoc query. The lengthy data transformation cycles led to the redesigning of the BI dashboard views which created multiple sources of truth. This in turn puts immense pressure on the enterprise's analytics team to give real-time relevant insights.

Leni was deployed to redefine the analytics workflow for the unit and reduce the time to insight. Business users were now able to directly pose their queries to Leni for getting real-time insights with simplified narratives on revenue, new business, losses, premiums, and RPC and retention rates across different product groups, segments, and lines of business to help them make effective business decisions. This drastically decreased their dependence on analytics teams and made their process more streamlined with real-time analysis.

For example, questions like why has revenue changed in the last month for the commercial segment or what would be the commercial segment revenue in the next 6 months, which typically took 2-3 weeks of turnaround time were now being answered by Leni in seconds using advanced analytics models. The team could also understand the steps of the analysis using Leni's explainability feature to gain more trust and confidence in the results.

Are LTIs insurance clients currently using LENI? Has the application been piloted? How did that go? Is it now live at any insurers? Please provide some clients using, and description of the kind(s) or entity or entities.

Yes. One of the leading P&C insurers is using Leni as the cognitive insights platform for their underwriting business. The pilot ran for 3 months with Leni & the client was able to achieve a 10X improvement in time to insights. The pilot was well-received, and discussions are underway to roll it out for other use cases.