Apr 20, 2023
Learn how credolab's solutions evaluate risk, detect fraud and optimise marketing in four categories of embedded finance, using behavioural data analytics.
This listicle is adapted from our latest series on A Beginner’s Guide to EmFi. To learn more about the different categories of EmFi used across industries and understand the main barriers to implementing this new model, check out the full article here.
Embedded finance (EmFi) has emerged as a major trend that may transform financial services. By disrupting the traditional financial industry, EmFi offers an innovative approach by integrating financial services into non-financial products or services using enablers such as Banking-as-a-Service (BaaS) or Application Programming Interfaces (APIs). Thus, it provides businesses with access to sensitive financial information and gains customer insights. This creates new opportunities for businesses to provide financial services that are more convenient and accessible for customers, making transactions seamless and efficient.
According to a report by Dealroom, EmFi can open up opportunities beyond the combined value of all fintech startups and top world banks and insurers. It is also expected to grow its worth to more than $7 trillion by 2030. As more consumers adopt digital payments, digital-first business models emerge, along with APIs and open banking, which leads to the industry significantly growing over the next few years.
EmFi comes with potential risks, as with any new technology and increased convenience. These include customer and security risks and the potential for fraud. In order to address these challenges, businesses such as credolab have developed solutions to mitigate risk, detect fraud, and optimise marketing spend. Credolab's behavioural data analytics platform uses machine learning (ML) to identify behavioural patterns and calculate insights from smartphone and web metadata.
This article will explore how businesses can level up with credolab solutions in the four categories of EmFi: embedded banking, payments, insurance, and investments.
Embedded banking provides a new way of integration allowing non-financial businesses to offer their users banking products like a savings account or a branded card, such as linked accounts and debit or credit cards. For example, integrating traditional banking services like checking accounts and debit cards with non-financial businesses like retailers. With the rise of digital and online transactions, the mobility of money has also increased, especially with embedded banking at play.
The shift to a digital landscape has increased the difficulty of avoiding scams or stopping potential fraudulent transactions. Although advantageous to customers and businesses, the ease of digital and online transactions also benefits fraudsters who can now exploit the vulnerabilities of Banking-as-a-Service (BaaS) solutions and gain access to sensitive financial information. BaaS solutions offer the duality of financial and non-financial services, such as lending licences to other clients. In contrast to a digital version of the bank, BaaS allows banks to expand their services and open new revenue streams.
However, BaaS has its drawbacks, including its difficulty in implementing as banks have to de-couple their legacy banking system and turn it into a fully digital, composable or modular technology stack or build an entirely new one. This overhaul leads to the challenge of predicting whether an applicant will commit fraud. The best way to tackle this is by adopting a proactive approach, especially since first-party fraud detection determines the likelihood of a payment default during the loan application process. This is where credolab's solutions for early fraud detection using ML algorithms come into play.
Credolab offers solutions for early fraud detection using ML algorithms, leading to the seamless reduction of first-party and third-party fraud at the onboarding stage. These ML algorithms identify behavioural patterns from about 70,000 privacy-consented and permissioned data points collected and transformed into over 10 million behavioural features. As a result, tools such as device velocity checks, IP address velocity checks, and fraud scoring can be turned into potential red flags able to detect first-party fraud, data manipulation, and device and behavioural anomalies.
Furthermore, credolab leverages ML to process large data sets and create models to find anomalous patterns between user behaviour and the likelihood of suspicious actions. Using signals built on top of the smartphone and web metadata makes it possible to collect and then convert raw data into granular customer insights. These insights, created in conjunction with TransUnion’s TruValidate product, will help flag potential scammers for each applicant in a secure, non-intrusive, frictionless, and real-time manner.
As a global partner of the TruValidate product, credolab offers a one-stop-shop solution for combating fraud and assessing risk during onboarding. Businesses can make more informed decisions by combining credolab's checks and scores with internal checks. They can even automatically blacklist or reject IDs once they have been confirmed fraud cases using analysed data from ML algorithms detecting fraudulent signals or alarms.
Credolab's ML algorithms help discover behavioural patterns likely to be linked to suspicious or fraudulent behaviours. In conjunction with TransUnion's TruValidate product, these algorithms enable more informed fraud decisions to be made by transforming signals into rich insights.
Some of these signals used include:
The above-identified signals and flags can be turned into granular insights for businesses to assess and mitigate potential fraud risks.
As addressed in A Beginner’s Guide to EmFi: Part 3 How to Leverage credolab’s Solutions, these identified red flags are combined with user-level, granular insights and complemented with:
Using credolab's solutions, businesses can level up their fraud detection and strategy to achieve optimal results, such as up to 23.9% reduced fraud costs and up to 21% decreased first payment default. In addition, with the rise of embedded finance, credolab's solutions provide a proactive approach to detecting fraud, especially in embedded banking, before it happens, available to all businesses, including non-lenders and non-financial institutions.
Embedded lending is a concept that is gaining traction in the financial industry. It is a growing trend involving integrating loan and instalment payments into non-financial products and services to improve loan approvals' accuracy. In the absence of credit histories, assessing risks and maintaining approval accuracy can be difficult, and it becomes a challenge for traditional lenders. This is because credit histories were essential for lenders to accurately determine a borrower's risk level. Consequently, traditional risk assessments tend to exclude those who are unbanked or underserved and who are more likely to receive loan rejection. This method leaves out many people who could benefit from financial services and only contributes to widening financial exclusion.
One of the most common examples of embedded lending is Buy Now, Pay Later (BNPL) services. According to a report from the Consumer Financial Protection Bureau (CFPB), the number of BNPL loans issued to consumers in the United States (US) grew from 16.8 million in 2019 to 180 million in 2021. Furthermore, BNPL's borrower base consists primarily of younger borrowers and leans more toward Millennials than Gen Z. They provide more favourable lending options as these services are often available to those with no credit history, low credit scores, or access to traditional financial services. This alternative option helps improve customer risk assessment and, as a result, opens doors to new customers. However, customers, merchants, and providers of BNPLs face risks such as being unable to repay the full amount, which can damage their finances and reputations.
Proper security strategies, such as rule-based risk assessment and real-time data enrichment, must be implemented to overcome these challenges and ensure all parties' safety.
Credolab offers solutions that improve credit risk assessment, even for those without credit bureau histories. Going beyond a rule-based risk assessment, credolab's behavioural data analytics platform leverages ML algorithms to identify behavioural patterns from smartphone and web metadata, such as using a single device, browser or IP address for repetitive loan applications and analysing the frequency of applicants changing key information such as income or address. These behaviours are converted into granular insights to assess risk and predict the probability of early default. After all, this assessment is the first step in building an underwriting and fraud strategy that leverages multiple data sources for any embedded lender, especially BNPL providers.
By implementing a robust risk detection strategy, credolab's risk solution can help businesses reduce non-payment risk, prevent fraud, and improve customer experience and satisfaction. Credolab's insights, such as device, browser, or IP address usage, combined with BNPL providers' existing (and, at times, rather traditional) strategies, can determine a borrower's risk level and legitimacy. This is done by processing raw behavioural data into scores that identify the correlation with micro-behavioural patterns similar to applicants who are identified as fraudulent or risky. Furthermore, the emergence of new data sources using ML and Artificial Intelligence (AI) algorithms allows traditional and alternative data to create more predictive scoring models.
With its unique source of smartphone behavioural metadata, credolab is also a trusted partner of TransUnion, which has integrated the credoSDK into the TransUnion Digital Onboarding solution, tailored specifically for BNPL customers. Combining these two technologies provides a seamless, end-to-end credit risk assessment, fraud assessment, compliance assessment, identity verification service, and payment facilitation across multiple touchpoints powered by the power of device data.
Any customer's first experience with a financial institution sets the tone for their overall satisfaction. In most cases, the first encounter occurs during the onboarding process. With TransUnion's Digital Onboarding and credolab's solutions, financial institutions can simplify their application process, minimise the need for multiple integrations, and conserve time and resources for IT teams.
With a secure and multi-functional API system, this solution enables an easy signup process, including ID document scanning, OCR, and data verification against TransUnion's extensive identity data sets. As a result, the application form is pre-filled to prevent human error and enhance the user experience.
Identity verification ensures that the person applying for a loan or opening a bank account is whom they claim to be. Customer identity verification refers to the validation process of an individual's information and identity.
TransUnion enhances identity verification by evaluating the applicant's consumer identity with background checks on the applicant's consumer data such as consumer email, phone, device SIM swaps, number portability, anti-money laundering screenings and geolocation checks. Furthermore, using Truvalidate's identity verification for identity insights also helps detect fraudulent or stolen identities through a robust view of identity risk, real-time fraud alerts, and contextual alerts for SSN verification on top of consumer data.
As a result, the applicant's consumer identity can be securely verified, and identity fraud can be simultaneously assessed to expose any fraud risks with the gathered identity insights.
Transunion evaluates consumer risk transparently in the background with its thorough identity verification. By integrating CredoSDK into TransUnion's digital onboarding solution, lenders can conduct alternative credit checks alongside affordability checks and avoid risky lending procedures.
The easy-to-use application is finalised by verifying all information and bringing consumers one step closer to their purchase. Ultimately, overall risk evaluation is improved as credoSDK work in collaboration with TransUnion's solution to assess risk and reduce fraud risk with the following rules:
Digital identity verification expands the concept of identity verification to today's remote world using processes such as biometric authentication, facial recognition and digital identity document verification. As a result, companies, governments, and financial institutions can securely and seamlessly verify an individual's identity using these digital verifications.
After verification, customers can pay the first instalment using their credit cards or bank accounts. Once the account is completed, future access is seamless using biometric authentication, which refers to verifying the authenticity of a person, where it compares the applicant's individual characteristics to their biometric "template". This authentication further reduces fraud risks like account takeover fraud by using unique physical characteristics such as fingerprints or face IDs to verify one's identity. Furthermore, this will increase friction-right consumer engagement, resulting in higher conversion rates and reduced fraud and credit risks.
Using credolab's solutions, businesses can take their results to the next level to generate an alternative risk score for all mobile and web applicants and achieve optimal results, such as up to 21.9% reduced risk costs and up to 32% increased approval rates. Moreover, as embedded lending enables access to new customers and offers more favourable lending options, credolab's solutions provide a proactive approach to assessing borrowers' risk levels for non-lenders and non-financial institutions.
Finding a credit card and entering its information can lead to customers disregarding a digital purchase if they do not have their card on hand. With embedded payments, individuals can save their preferred payment method for future online purchases.
On the other hand, businesses are beginning to see the benefits of embedding these processes across the whole value chain and controlling the entire revenue stream. The embedded payment process facilitates checkout, optimises time, improves customer experiences, and reduces the risk of cart abandonment. Fast embedded payment is already available in many apps, especially delivery and transportation apps.
It is, however, a top target for fraudsters, causing a growing concern for businesses and consumers alike. According to the latest estimates, criminal fraud involving payment cards will cost $49.32 billion out of the total payment card volume of $79 trillion by 2030.
One of the most common forms of fraud affecting businesses worldwide is chargeback fraud, also known as friendly fraud, which occurs when people falsely deny using their credit or debit cards in order to receive a refund. Transactions using cryptocurrencies like bitcoin are not subject to chargebacks but are still vulnerable to other types of fraud, such as transaction fraud. This occurs when stolen payment cards or data are used for unauthorised transactions. In crypto, fraudsters use tactics like investment schemes where they impersonate businesses or government agencies and convince people to buy cryptocurrency, which they then steal for investment. Account takeover fraud is another tactic where fraudsters steal private information by intercepting Wi-Fi signals in public locations. In 2022, hundreds of user account on Crypto.com were hacked, leading to stolen funds.
According to a Federal Trade Commission report, cryptocurrency scammers have stolen over $1 billion since 2021. Thankfully, chargebacks and transaction fraud can be stopped in real-time. Alternative data makes it possible to identify legitimate customers and suspicious activity in real-time, minimising fraud by analysing users' login behaviours. This applies to all embedded payment options, from credit cards to crypto.
"Our solution leverages Machine Learning (ML) algorithms to detect early fraud by identifying behavioural patterns. By converting anonymous data stored on smartphones into quantitative figures, we can assess an individual's likelihood of defaulting and committing fraud. Using a combination of our proprietary smartphone device ID and first-party metadata, we can generate a real-time fraud score for each applicant in a secure and non-invasive manner, with their consent." By Michele Tucci, MD Americas & Chief Strategy Officer at credolab
The failure to detect fraud during the loan application stage can often result in a default on a loan. Therefore, in order to prevent being blindsided by any cyber, risk or fraud attack, fraud prevention strategies and fraud detection tools have become an essential part of business protection. In addition to educating employees and offering regular data security training, businesses should also invest in building a scoring model that combines automation with various data sources to provide lenders with a holistic understanding of their customer's behaviour. With the emergence of new data sources, ML and AI, combining traditional and alternative data to create a scoring model has become possible. Combining these concepts also contributes to a better understanding of human behaviours, leading to more accurate and reliable models.
By working with credolab, businesses can collect specific information to perform checks for their fraud strategy, securing an effective user verification while exposing possible fraudulent behaviours. For example, these insights from smartphone and web metadata can:
With these insights collected, businesses can develop a better understanding of the users' behaviours while at the same time keeping the identity of each user protected and avoiding the kind of cyber crimes that attempt to steal customer information and commit fraud. For example, businesses can detect potential risks, assess how often a particular device or IP address attempts to access the network and limit the risk of a data breach before it occurs, ensuring data security and improved privacy for their customers.
Embedded insurance provides customers with the opportunity to buy insurance while making a purchase. It empowers customers to have the ability to purchase more affordable and personalised insurance efficiently. Some examples of natively embedded insurance or otherwise known as "Insurance as a Service (IaaS), include:
Through integrating insurance on consumer platforms like Airbnb, customers can enjoy customised coverage without going through a third party. This solution can benefit businesses and insurers, which offers branding opportunities, revenue growth, and long-term customer loyalty. Furthermore, their customers can benefit from reduced costs and contract flexibility. For instance, the Singapore car marketplace Carro deploys embedded finance strategies to create a comprehensive brand offering various services like contactless online purchases, usage-based insurance, and car rental subscriptions.
As the market evolves, embedded insurance faces several challenges. Competition has intensified as more players enter the industry, increasing advertising and lead generation costs. Additionally, the industry has been slow to adopt digital technologies that meet the needs of modern insurance buyers. To succeed in this crowded market, businesses must develop more personalised approaches to insurance to attract and retainer customers.
In order to improve underwriting efficiency, many insurers mistakenly believe automating manual processes is enough. In reality, this only alleviates superficial obstacles, and the old limitations remain. For modernisation to succeed, it requires more than improving aged systems but also adjusting mindsets and exploring completely new approaches.
Through behavioural data analytics, insurers can transform their relationship with policyholders from transactional to customer-centric. Understanding the customer's perspective throughout the entire process of selling a product is essential. With behavioural data analytics, rather than acquiring intermittent feedback through expensive research methods, immense volumes of information can be gathered and analysed from different sources, and consumer behaviours understood across various touchpoints can be understood.
With credolab, it is possible to understand customers better and offer an improved overall experience: services respond effectively, claims are processed instantly, and policy writing is done faster. Using its proprietary data modelling pipeline rooted in ML algorithms, credolab analyses over 70,000 privacy-consented data points from a smartphone device, converting customers' digital footprint into highly predictive scores tailor-made to each client. In the case of insurers, credolab can help analyse user behaviour, assess risk, assist insurers in underwriting, and offer opportunities to potential customers, allowing them to understand better, anticipate policy lapsations, and optimise cross-sale, upsell, and retention strategies.
The persistency ratio indicates the number of policyholders paying premiums for active plans. About 20% of all new customers fail to pay their first bill on time, especially if their credit card is not linked to the payment. When this happens, underwriting costs, brokerage fees, and other acquisition costs make the policy unprofitable without being paid for.
By working with credolab, businesses can leverage the alternative data sources provided to predict the probability that a new applicant will not make the first premium payment. These additional details allow insurers to understand their clients better and make more informed decisions regarding pricing, coverage terms, conditions, and risk levels.
The lapse ratio measures customers who have yet to renew their policies with an insurance company. Insurers closely monitor the ratio because it helps them determine how efficiently they retain their customers and earn profits. Approximately 15% of customers miss payments between months 6 and 8, which is when the lapse ratio becomes an issue. In these cases, even if the insurer stops premium coverage, it typically loses money if the policy has yet to break even.
To address this issue, credolab can help insurers to reduce the number of delinquent customers approved at the origination/application stage. This is achieved by using smartphone device metadata to arrive at a fraud score for each applicant in a non-intrusive and secure manner and in real time. This way, credolab can help insurers keep a healthy policy lapsation ratio by predicting which applicants' policies will most likely stop paying before months 10 and 11.
A healthy fraud control system is critical to the profitability and ROI of insurers. In insurance, fraud can occur from both sides: the seller can sell policies from nonexistent businesses in order to generate a higher commission, and the buyer can make false claims, such as falsifying accidents or theft. Therefore, insurers usually charge their clients higher premiums to cover this cost.
Digital fraud is among the most prevalent in the insurance industry. According to a TransUnion report, insurance had the fifth-highest number of digital fraud attempts from 2020 to 2021, only behind gaming, travel and leisure, telecommunications and financial services.
Credolab mitigates this problem by detecting fraud at the origination stage using behavioural data insights from alternative data. As a result of their ability to identify mobile devices (their brand, operating systems, and models) and anomalous behaviours (such as autofill forms, VPNs, or TOR), these insights prevent malicious subscriptions. In addition, the application can be used to discover the following:
As a result of the above factors, an effective onboarding process can be safeguarded, and fraudulent behaviour can be detected.
In the business world, customer acquisition costs are a common problem. Compared with customer retention, most companies spend five times more on acquisition than on retention on average, while insurance companies spend seven to nine times more on acquisition than customer retention. The high cost incurred by insurance businesses is largely due to the outdated acquisition methods that most are still adopting. For instance, TV ads cost insurance businesses hundreds of millions of dollars a year but do not directly benefit customers.
Credolab helps lower the average acquisition cost by increasing the acceptance rate of applicants. Insurers today tend to turn away applicants (especially in emerging markets) because they lack the data to make an informed decision. By using behavioural data insights from alternative data, credolab improves the allocation of leads by calculating their probability of accepting an insurance offer. Furthermore, credolab provides a score for 100% of incoming requests, which reduces data asymmetry when making a decision and increases approval rates.
The conversion rate increases if leads are segmented correctly, personalised, and directed toward the insurance the customer will most likely accept. By leveraging credolab's behavioural data analytics platform, insurers can better understand their customers and optimise their customer acquisition costs.
EmFi is transforming the financial services industry, providing innovative approaches to integrate financial services into non-financial products or services. However, with increased convenience and accessibility, businesses should also consider the potential challenges and risks of EmFi, specifically in the four categories - embedded banking, lending, payments and insurance. Fortunately, businesses can mitigate these risks using behavioural data analytics platforms like credolab's and proactively level up their fraud detection, risk assessment and marketing optimisation.
Credolab's solutions offer businesses a unique advantage by leveraging ML algorithms and behavioural analytics using behavioural patterns to generate insights for detecting early fraud and assessing customer risk. Moreover, by combining credolab's solutions with internal checks and TransUnion's Truvalidate and Digital Onboarding Process, businesses can:
Credolab's solutions, available to all businesses, including non-lenders and non-financial institutions, for EmFi, help businesses level up and offer more financial services to a broader range of customers to create an increasingly inclusive world.
Access data insights solutions that deliver growth - Fraud detection | Credit scoring | Marketing segmentation. Helps you say "YES" more confidently to more customers!
Learn more about credolab's products and possibilities with our features through our Blog section, and feel free to share our content with your team!
March 30, 2023
How Embedded Scoring Can Help Financial Institutions: A Detailed Overview
March 16, 2023
A Beginner’s Guide to EmFi: Part 3 How to Leverage credolab’s Solutions