Q&A with Martina King at Featurespace
Featurespace™ is dedicated to making the world a safer place to transact for its clients and customers. They have developed market-leading solutions to fight enterprise fraud, scams, and financial crime using Adaptive Behavioral Analytics and Automated Deep Behavioral Networks, both of which are available through the ARIC™ Risk Hub, Featurespace’s proprietary real-time machine learning platform that risk scores events to prevent fraud, scams, and financial crime.
Liz Beggs, Chief Operating Officer of Payments Consulting Network, sat down with Martina King, CEO of Featurespace™, to learn more about the company, what their offerings are and some of the exciting projects they are currently involved in. Martina joined Featurespace™ in 2012. Formerly the Managing Director of augmented reality company Aurasma, she has an extensive career in media and technology, including MD of Yahoo! Europe and Capital Radio. Under King’s leadership, Featurespace has become the world’s leading provider of Enterprise Financial Crime prevention software, risk scoring events and preventing fraud and financial crime in more than 180 countries. The company has also raised more than £100 million in funding, supporting its ongoing expansion and ability to protect more than 30 major global financial institutions, including HSBC, TSYS, Worldpay, NatWest Group, Danske Bank, BBVA and Barclays PLC. Since 2018, she’s been honored by PaymentsSource as one of its Most Influential Women in Payments and has previously been named “Top 40 Powerful Women in Tech” by Silicon Republic. Martina is also President of the London Chamber of Commerce.
LB: Please provide a brief high-level overview of your business
MK: Featurespace™ is the world leader in Enterprise Financial Crime prevention for fraud, scams, and money laundering. Featurespace invented Adaptive Behavioral Analytics and Automated Deep Behavioral Networks, both of which are available through the ARIC™ Risk Hub, Featurespace’s proprietary real-time machine learning platform that risk scores events to prevent fraud, scams, and financial crime.
Featurespace was created by Cambridge University Professor, Bill Fitzgerald, and his PhD student, Dave Excell, at the forefront of two academic fields: Data Science and Computer Science. Featurespace’s Adaptive Behavioral Analytics technology, invented by Bill and Dave, had its premiere in 2008, when online gaming company Betfair asked Featurespace to build the first system to fight the increasing prevalence of fraud on their rapidly growing platform. The world’s first Adaptive Behavioral Analytics engine – the ARIC™ Risk Hub – was built to solve this commercial challenge.
Today, ARIC Risk Hub is relied on to catch new fraud and financial crime by identifying suspicious activity in real time at more than 70 major global financial institutions.
Publicly announced customers of Featurespace include banks (HSBC, NatWest Group, BBVA, Danske Bank, AKbank), some of the world’s largest acquirers and payment processors (TSYS, FIS Worldpay, Marqeta, GPS, eftpos), and next-generation payments/banking-as-a-platform services (Soldo, Railsr).
The business has strong momentum, and we are experiencing record international growth across Europe, the Americas and Asia Pacific. As a result of the performance of ARIC Risk Hub, we have high customer and revenue retention.
Alongside significant business progress, Featurespace has received global recognition from reputable industry organisations such as Aite-Novarica Group, Forrester, FStech, IDC, Celent, Cards and Payments, and Finovate. In 2021, we were presented with the Queen’s Awards for Enterprise – International Trade, our second time to have received this award.
LB: What types of fraud management and financial crime services do you offer companies in Australia?
MK: With our six solutions, there is a myriad of use cases for banks and Financial Institutions to select within our ARIC™ Risk Hub. Most in demand deployments are used for solving scams, payment and card fraud, as well as money laundering. A 75% catch rate for card issuers, using models and authorisation stream data only, provides banks with the most secure way to protect their own customers’ payments and reduce their own exposure.
For acquirers, issuers and payment processors, we supply multi-tenanted technology which enables them to deploy market leading fraud and financial crime protection that their own clients then utilise. This route to market covers issuing and merchant acquiring and again outperforms all others.
Increasingly, customers are seeking a truly adaptive platform and flexible technology that gives their own data science and analytics teams the ability to deploy and manage their own analytics. ARIC Risk Hub’s inherent adaptability enables financial institutions to tackle current and future threats more effectively and efficiently, easily orchestrating data from across their organisation, to outsmart criminal networks.
Our partnership approach is the best way to ensure customers achieve their goals and get full value from our technology.
LB: Are there any industry sectors or client types that you focus on for fraud managment and financial crime services?
MK: The industry sector we serve is banks and other financial institutions with direct customer relationships, as well as acquirers, issuers and payment processors.
LB: What do you see as your key strengths with respect to your solutions?
MK: Featurespace is the first company to profile both genuine and fraudulent behavior to better identify fraudulent activity in real time; our proprietary innovation is known as Adaptive Behavioral Analytics.
Featurespace was set up in Cambridge, by people at the forefront of machine learning and our heritage is rooted in research and science. The application of the very best technological thinking is suffused in our DNA and is applied to everything we do. This is why our real-time self-learning technology is proven to outperform all others.
Featurespace’s latest innovation, Automated Deep Behavioral Networks, gives our customers a novel deep learning implementation for even greater fraud protection. Purposely designed for the Card and Payments industry, Automated Deep Behavioral Networks improves fraud detection across the board by >30% over already industry leading models and is the latest step in our innovation journey.
ARIC Risk Hub is uniquely positioned to help our customers achieve market leading results based on four promises:
- Speed: real time protection prevents fraud before it happens. Quick to install, ARIC Risk Hub delivers low transaction latency, high event throughput, and reduced response, investigation, and handling time.
- Accuracy: ARIC Risk Hub delivers improved fraud detection rates for both value and volume, reduces false positives and lowers business overheads.
- Adaptability: adaptive models ensure accuracy is maintained over time irrespective of the changes in customer behavior. Fraud strategists and customer data scientists are enabled to update strategies to suite changing business needs and better serve customers.
- Simplicity: ease of implementation into customers’ existing strategy. Access to the very best 3rd party services ensures customers know why a model is making a decision without bias and can stand up to the most stringent model governance.
LB: What do you see as your key differentiators with respect to your solutions?
MK: Featurespace’s main differentiators vis-à-vis our competitors are:
- Real-time transaction monitoring with extremely high throughput
- Real-time behavioral profiling used by rules and models
- Auto-retraining models based on fraud labels (confirmation events or analyst reviews), keeping the solution’s Total Cost of Ownership low
- Explainable analytics
- Consultancy services with fraud, data science and analytical subject matter experts who have run fraud shops and successfully deployed compliance models within several markets
LB: What were your key achievements over the last 12 months?
MK: We were awarded two new patents recently for our innovative technology that will help fight global financial crimes. The first patent is for our Fragmentation Engine, an innovative solution that enables ultra-low latency real-time scoring capabilities. The second patent is for our Sandbox Layered State that enables the addition of new behavioral elements to existing behavioral profiles with zero impact on real-time processing flow.
Our latest patents are intended for use by regulated financial services, payment service providers (PSPs), banks and businesses, in the international fight against fraud, scams, and financial crime. Our patent technology is designed with legal and regulatory frameworks in mind, such as the EU’s Artificial Intelligence Act and the UK Bank of England’s Artificial Intelligence and Machine Learning regulation.
New clients who have put their trust in our technology include Railsr, Soldo, VGW, Equals, Monavate, and BBVA.
LB: I have read that Featurespace is currently prototyping its specific services in line with the PETS (Privacy Enhancing Technologies) challenge guidelines at the moment and will submit the finished solution in advance of the official deadline on 24 January 2023. Can you explain this a little further?
MK: The PETs Prize Challenge is jointly hosted by the US and UK Governments. There are two tracks – financial crime (which we are participating in) and pandemic response. In both tracks, there is a serious societal problem where the pooling of individual sensitive citizens data would result in better outcomes for those citizens and offer a safer world to live in. However, sharing that data is in tension with individual rights to privacy, anonymity, and control over the person’s data.
As a result, the challenge has called for a technological solution, where the necessary insights to make the world safer can be drawn from an individual’s data and shared, but the underlying data itself remains entirely private. This called for privacy-preserving adaptations to the idea of federated learning, where parts of a machine learning model are trained on machines belonging to different organisations.
The Governments organised the challenge into 3 Phases:
Phase 1 was a call for proposals, in which 76 teams from around the world submitted technical whitepapers outlining a variety of different approaches to solving the problem. Featurespace’s proposal was one of 12 proposals awarded a Phase 1 prize and we were invited to participate in Phase 2 with full funding from Innovate UK.
Phase 2 called on us to build a prototype and submit it on 24th January. A key element of the Phase 2 prototype was to show that the privacy-preserving federated learning system was highly effective at spotting financial crime. Our solution scored top in the UK leaderboard. Our prototype has now been selected as a finalist for Phase 3, which will see Government-sponsored red teams (i.e., hackers) try and break the privacy guarantees provided by our prototype. Surviving this attack stage will validate the guarantees we provide.
The results of Phase 3 will be announced at the Summit for Democracy on 29th March.
LB: Featurespace’s next gen AI technology is set to be showcased at the second Summit for Democracy to be convened by President Joe Biden (USA) in March 2023. What would Featurespace do with the money if it is successful in winning one of the prizes? In addition, what do Featurespace hope to get out of this Summit?
MK: If we win one of the prizes for Phases 2 & 3, we will reinvest that prize money into our Innovation Portfolio. Privacy enhancing technologies is only one of the programs of innovation that we are pursuing at the moment. Innovation is a continual process of generating ideas and propagating them into products. In high tech industries like ours, this process can never stop!
LB: What were your key findings from your recent report on The State of Fraud and Financial Crime in the US and what do you believe Financial Institutions can do to protect themselves in the future?
MK: The report paints a picture of a sector trapped between increasing financial crime – especially scams – and the perception that any solution will bring added complexity and compliance headaches.
Key findings include:
- Two-thirds of Financial Institutions report an increase in volume and dollar cost of fraudulent transactions.
- Fraud is on the rise across all payment types, with credit cards most commonly targeted.
- Scams make up a quarter of fraudulent transactions, with average losses of over $100M last year across the surveyed financial institutions.
- More than half (58%) of executives worry that no solution can match the sophistication of attacks.
- The industry now acknowledges machine learning as the preeminent method for solving fraud, scams, and financial crime.
Two-thirds (66%) of respondents admitted that “complex regulatory requirements” were holding them back from trying new technology options to counter crime.
Moreover, 59% of executives said they were adopting a ‘wait and see’ approach until newer technologies are ‘widely accepted’ or ‘well-developed’ before considering them as alternatives to older, outdated mitigation options. Those who launched innovative solutions before others saw the lowest fraud rates, while ‘fast followers’ saw significantly lower fraud rates.
The data – alongside our own experience – shows there’s an appetite for more innovative solutions able to address the ever-increasing challenges posed, yet it appears some institutions continue to wait before taking the leap and benefiting from the significantly reduced fraud losses promised to smart thinking, first movers.
Fears of compliance or technology complexity appear to be holding institutions back from employing more effective machine learning, fraud detection platforms that are simple, battle tested, and better equipped to minimise institutional losses.
Our hope is that this report has opened an industry wide conversation, and we will design future iterations of the survey to build a deeper understanding of the challenges we face, and what we must do collectively to outsmart the criminal. The one thing we can’t do is nothing.
LB: What other fraud management and/or financial crime innovations do you have on your product/service roadmap for the next 12 months?
MK:
- Helping firms combat scams both on the inbound and outbound payments front.
- Helping firms to optimise their spend with external data providers.
LB: What industry changes or trends to do you see occurring over the next 2-3 years that will have a major impact on your business and/or your clients?
MK:
- There has been a huge amount of coverage around financial stability with costs increasing in all areas: gas, electric and food etc. The continued conflict between Ukraine and Russia is also impacting the financial status of the world. As such, companies are looking internally at their operational costs and identifying where savings can be made with the goal of doing more with less.
- At the same time customer experience and acceptance rates of transactions are also a key area of focus in all regions. This has always been a challenge with banks weighing up risk against reward, which is a continuous battle. Previous fraud approaches with rules-only are simply not proving effective and this is why machine learning is not becoming a nice to have but a ‘Must-have!’
- Speed and the ability to respond to change in an agile way are also important factors for Financial Institutions to avoid increased fraud losses, regulatory fines, or customer attrition due to the emergence of new fraud typologies or change in government regulations. However, legacy systems do not always make this easy. Therefore, moving to Cloud-based solutions is becoming more and more attractive.
- Scams have been a big part of fraud and have exploded since the pandemic and for the foreseeable future this does not seem to be changing. Fraudsters are exploiting people by persuading them to take action through convincing them that they are in fact the bank, police or government for example.
- This type of attack leans on the human nature of people respecting authority and wanting to cooperate. Despite the huge number of media articles about scams, they continue to probe and have success.
- Investment and romance scams are typically longer running scams where the fraudster builds up a rapport and trust over time and then starts to apply pressure, so the victim feels they need to send the fraudster money. Existing education doesn’t seem to be working and drastic changes are needed to halt the tide.
- PSD3 (and other global interpretations) will focus on plugging the gaps left by PSD2, and in turn will force the criminals to resort to other methods.
LB: What key criteria or features should a business consider when evaluating fraud and financial crime solutions?
MK:
- Cloud native solutions are becoming the preference that we see our customers are looking for from their fraud and financial crime solutions. It is becoming a fundamental requirement when selecting vendors who are able to offer cost-effective hosting services, with a lower total cost of ownership.
- Fast and seamless deployment of a solution is critical to technical teams. Most businesses are always dealing with multiple challenges, relying on the same people within their company so vendors who can help with the heavy lifting of providing their services will prove to be an attractive proposition.
- Results! Having a proven track record with customer endorsements and recommendations is a key factor.
- Enterprise solutions being able to solve multiple use cases with one deployment.
- API readiness – how easy it is to integrate to internal and potentially other external systems.
Author: Liz Beggs, Chief Operating Officer, Melbourne, Payments Consulting Network
Liz Beggs is a versatile and results focussed leader with over 25 years of multi-industry experience in consulting and management. She delivers value add, innovative solutions to internal and external clients with subject matter expertise in commercial cards and payments technology systems, drawing on her broad commercial GM level experience in corporate services and product management.
***
Featurespace is a member of our Fraud and Chargeback Management Panel.
If you found this article helpful and would be interested in reading similar articles, please subscribe to our newsletter.
To get notified of our latest posts, follow Payments Consulting Network company LinkedIn page and click on the bell icon at the top right section of our company profile.