Take appropriate actions while your customer is still on your site and avoid inappropriate behavior in complex online environments.
Online Game Fraud Detection and Prediction
Fraud in online games mostly consists of the illegitimate acquisition of assets that are normally purchased or won via game play. tazi anti-fraud solutions use smart algorithms based on the continuous collection of real-time data from online games to reduce game fraud.
Online Game Recommendation
tazi can monitor player activity and recommend other players as potential partners or opponents. It also recommends possible game elements that could increase the likelihood of player retention.
As regulations, supply and demand and customers themselves change constantly, so do the customer purchase patterns. Thanks to tazi’s online machine learning infrastructure and rapid adaptation of machine learning models, purchase prediction is estimated more accurately.
Next Best Content or Campaign Offer
tazi analyses user, item, content, context and all other available information. It identifies the important factors that affect acceptance of offers by the customers.
tazi processes relevant and heterogeneous data produced by customers online. In order to determine the next best offer, it learns from all customers' transactions as they happen.
Through tazi's friendly user interface, campaign experts can get insights on machine learning models, and, if required, update their models based on their knowledge. Their understanding of the models and their performance enables online campaign updates.
Classifies the behavior of streaming data as normal or anomalous in real time, with as little user feedback as possible.
Online Multiplayer Gaming: For an online game company with millions of monthly users, tazi’s machine learning system identified fraud with 98% accuracy. Fraud patterns change constantly making it very difficult for moderators to catch. Thanks to tazi, having a 44% reduction in false positive rate than the existing rule-based system, 2.4 times more fraud cases were caught without having to hire new moderators.
Banking: tazi processes network activity logs in real time and detects internal or external attacks on critical points within the banking network. tazi understands user activity patterns in online banking transactions, ATM usage, etc. tazi identifies anomalies in customer behavior in real time, understands the root cause, and suggests a solution with minimal customer inconvenience.
Insurance: Based on customer and claim data collected, tazi identifies if there are anomalies or patterns in the claims.
Call Center: tazi reviews call center data, such as call lengths and the number of calls with the same customer in a certain interval, and notices that a newly hired call center assistant keeps long calls, which usually end up with a bad rating from customers.
Internet of Things
IoT security, service recommendation, device and service monitoring.
Device/Service Anomaly Detection and Prediction
tazi sets a baseline for normalcy from data regularly collected from devices/services, and detects and predicts anomalies in their behaviour.
Regulations compliance requires the factory to emit only a certain amount of CO over a certain time interval, failure resulting in a fine. tazi predicts if the emissions will be beyond acceptable at a certain time in the future and warns the operator.
Based on the services available on hundreds of IoT devices, tazi recommends the right service for a specific user's needs, based on other users’ as well as the services’ behavior.
tazi can compute the relationships between thousands of sensors online.
tazi and tazi can identify the most important factors contributing to failure in the future so as to allow the operator to prevent such failure.
Critical Device and Service Identification
With so many devices in the IoT, one of them is likely to fail at any time.
Based on statistics gathered from the devices’ usage under different operating and network conditions, tazi predicts which devices are more likely to fail and are critical, so that those devices can be backed up with an alternative.
Learns customer preferences based on their behavior in real time, enabling dynamic recommendations.
Mobile Operator product recommendations: The task was to recommend the best offer from a portfolio of products to the existing customers. The client previously used a rule-based recommendation system. In the tests, tazi was able to process millions of transactions from millions of end-customers, performing 10x more accurately than the existing system. tazi was also able to intuitively present the reasons behind its recommendations, giving the client a better understanding of end-customer behaviour.
Insurance: tazi allows the client to offer the insurance policies best suited to their customers' needs. It can also determine when and how product offers should be made to customers so that they are more likely to accept them.
tazi is coming soon...
Banking & Insurance
Fast and accurate data analytics and insights for customer retention, offer management and fraud detection.
Customer Experience Optimization
tazi processes user activity patterns when users login to the online banking system, make online transactions, or use the ATMs or bank tellers. It produces real time and customized product offers while the customer is still on the company web site.
tazi processes customer behaviour data and predicts if a customer will churn. Through the analyses of churn models, preventative measures can be taken.
tazi allows offering the insurance policy which is best suited for the customer’s needs and is more likely to be accepted by the customer. It can also determine when and how should the product offers be made to the customers so that they are more likely to accept them.
tazi proccesses bank teller data, such as visit duration, the number of visits by the same customer in a certain interval, procedures performed during the visit, and more. Tellers or customers involved in unexpected or unusually long transactions are identified and appropriate warning mechanisms are invoked.
End User Behaviour
Customers living in a certain neighborhood of the city or of a certain segment start using their cards less than usual. tazi identifies this and produces alerts. The bank investigates and figures out that a rival company has started a new door-to-door marketing campaign in the region. The marketing department tailors a counter campaign to get some of the customers back.
Taking into account the past claims data and similar entities, tazi predicts the future probability of claims and also the amount for each insured entity and therefore the insurance cost can be computed.
Anomaly Detection, Prediction and Prevention
tazi processes the logs of different network devices together with personnel and customer activity. External or internal network or behavioral anomalies are discovered online as soon as they happen. Through root cause analysis of the anomalies, some can be predicted and prevented before they happen.
tazi monitors logs of the the transactions that take place while the insurance quotes and claims are produced and identifies the anomalies or problems and also their root causes so that appropriate actions can be taken for better customer service and claims processing.
Based on the customer and insurance claim data collected, tazi identifies if there are anomalies or patterns in the claims.
Catch up with numerous and continuously changing job requirements and applications.
Iterative Identification of Relevant CVs
An HR search company receives thousands of CVs when one of its clients posts a job advertisement. tazi evaluates and ranks each job application based on the job requirements and the applicants' CVs and histories. Now the HR specialist at the client company spends much less time identifying the candidates that need to be called in for an interview.
While the client company's HR specialist evaluates each CV for the current job ad, tazi learns from each evaluation and is able to re-compute the CVs' relevance.
tazi provides the client company's HR specialist with the underlying reasons behind its decisions, so that the HR specialist plans the interviews appropriately.
Pam posts her resume and searches for jobs with the keywords she is interested in. However, she does not get good results either due to poor choice of keywords or timing. tazi takes Pam's and previous applicants' CVs as input and recommends the jobs that Pam would like the most.
tazi screens job advertisements as they come and alerts Pam whenever it finds a good match.
tazi identifies the discrepancies between Pam's qualifications and those required by jobs relevant to her, helping Pam decide how she should strengthen her qualifications.