Kittipong Jirasukhanont/123RF
 

AI In Insurance Use Case #18: Tazi

  • 6 August 2019
  • Sam Mire

This interview is part of our new AI in Insurance series, where we interview the world's top thought leaders on the front lines of the intersections between AI and insurance.

In this interview, we speak with Zehra Cataltepe, co-founder and CEO of Tazi, to understand how her company is using AI to transform insurance, and what the future of the insurance industry holds.

1.  What’s the story behind Tazi? Why and how did you begin? 

ZC: What drives us is a dream of making AI usable by everyone.  We believe in democratizing it so that it becomes a useful technology like automobiles or electricity. When you consider business processes, life always changes. So, as an insurance company, the products that you offer, your customers' ability to purchase them, your competitors’ behavior, regulations — all of these change in time.

So, you have to keep adapting whatever you were doing to new changing conditions. You cannot just have one machine learning model or one rule set, and keep doing that forever. You need AI which can continuously learn, which can continuously adapt and which can do that adaptation without massive effort from its users. 

Established in 2015, Tazi builds on over 25 years of machine learning and large-scale systems experience in academia and industry. Our machine learning vision, shaped by real-life experience across telecom, retail, insurance, finance, and brand management, keeps business experts in the driver's seat while reducing the load on data science teams and increasing the ownership, accuracy, and value of the AI models.

2. Please describe your use case and how Tazi uses artificial intelligence:

ZC: Our team has 21 Machine Learning and Artificial Intelligence patents. Thanks to this expertise, Tazi automates time-consuming processes like data cleaning, feature generation, selection, building, and optimization of machine learning models.

Tazi’s Automated Machine Learning is understandable continuous machine learning from data and humans. It enables business domain experts to use machine learning to make predictions and take actions. It also helps data analysts and scientists for their daily model creation and deployment tasks.

Below is a sample set of our use cases:

Insurance Profitability & Growth Detection and Prediction

Tazi’s understandable continuous learning Auto-ML provides dynamic models providing new market insights (previously unknown sub-segments). This includes a 50 percent or greater improvement in detection of profitable & shrinking, loss & growing microsegments detection so that appropriate actions can be taken to increase revenue and reduce loss. Actuaries spend less time looking for variable combinations to define meaningful and actionable segments. 

Credit Risk Detection and Prediction

Heterogeneous data sources are integrated and processed using tazi's continuous ML technology. Past data on transactions of each customer, payment frequency, amount, currency, increase and decrease on the last payment ratios, and LCV are used to create features to detect and predict credit rating. We aim to show up to 75 percent accuracy for risk detection and 70 percent accuracy for predicting risk in the next quarter. 

Finance Non-Performing Loan Prediction

Based on features of the customer’s debt as well as interactions with the company, Tazi is able to predict who will pay with high accuracy. By just calling the top 60 percent scored customers, the agency can reach 90 percent of the customers who are likely to pay, reducing the call center costs and increasing customer satisfaction. 

3. Could you share a specific customer/user that benefits from what you offer? What has Tazi done for them?

ZC: In auto insurance, our customers are able to spend less time on figuring out profitable/lossy segments and more time making use of that data. Actuaries are able to create and select features and create machine learning models automatically at a fraction of time it would take to create a traditional actuarial model.

Our customers also love the fact that they can explore the model explanations in detail (up to instance level) and model performances continuously as they learn. Another advantage is the fact that the actuaries can create tens of models and maintain and monitor them, together with the other business units, all through an easy to use interface. 

4. What other AI use cases in Insurance are you excited about?

ZC:

Life Insurance – Churn Prediction

Tazi identified 89 percent of churning customers correctly with 75 percent accuracy. 

Auto Insurance – Claims Prediction

Tazi’s understandable and continuous models enable an 86 percent improvement in high claim detection. 

Health Insurance – Claims Prediction

We are able to identify who is likely to have a certain operation with six months with high accuracy. Use of this approach to identify who would be hospitalized due to for example diabetes or heart problems has enormous advantages, both in terms of cutting costs and also improving people's lives. 

Auto Insurance – Fraud Detection

Tazi can identify complex and evolving fraud cases accurately thanks to its ability to use different types of features, continuous learning and also the inclusion of the human fraud experts to accelerate and correct the machine learning algorithms' learning from data.

5. Where will Tazi be in five years?

ZC: We aim to enable not only insurance, finance or retail experts, but practically anyone who can benefit from the predictive abilities of the AI models, to create, understand, take actions with, update, and teach/correct those AI models. We aim to truly democratize AI by allowing anyone with a goal and data flows to make use of that data. We will become a multibillion-dollar company by increasing the total addressable market for automated machine learning from only data scientists and engineers to any ordinary person.

About Sam Mire

Sam is a Market Research Analyst at Disruptor Daily. He's a trained journalist with experience in the field of disruptive technology. He’s versed in the impact that blockchain technology is having on industries of today, from healthcare to cannabis. He’s written extensively on the individuals and companies shaping the future of tech, working directly with many of them to advance their vision. Sam is known for writing work that brings value to industry professionals and the generally curious – as well as an occasional smile to the face.

Comments

COMMUNITY