Vytas Kisielius, President, Adeptra

NewsBlast
Please tell our readers about Adeptra. How did you get started? How many people are in your organization? What markets do you have a presence in? What technologies do you develop?

Vytas Kisielius Adeptra was founded in London in 1996, and first hit the U.S. market in 1999 as Realcall, an e-commerce infrastructure product that helped web-based merchants create an interface between the Web site and the sales department. When the 'internet bubble' burst in 2000, we re-named ourselves Adeptra and re-architected our 2-way interactive alerting platform to enable high-ROI applications for enterprise-class (read "Fortune 100") clients. We're nearing 40 full-time staff in offices near London (Reading, Berkshire) and New York (Norwalk, Conn.) who focus on deploying Just in Time Response Solutions for Financial Services clients. We've developed both the alerting platform and the integration and reporting engines needed to turn a "cool technology" into client-facing applications such as fraud verification, payment collections, card activation, and retention marketing.

NB
How did Adeptra get involved in assisting in the fight against credit card fraud?

VK In 2001, we met some visionary managers in Chase and Fleet Bank's fraud departments who described an operational problem that we felt could be addressed with a Just in Time Response Solution approach. They agreed to pilot the Solution and were instrumental in our developing the fine points of the application to ensure it delivers an exceptionally high ROI. Our solutions are now being deployed by six of the top 10 card issuers in the United States, plus a growing list of banks in the United Kingdom.

NB
Can you give some examples of problems that banks are experiencing concerning credit card fraud?

VK Credit card fraud is a large and rapidly growing problem, exacerbated by the ease of "information theft" made possible by the Internet and online banking. Our clients are constantly trying to stay a step or two ahead of the fraudsters - who are frustratingly clever! Since the bank is on the hook for any fraudulent charges made beyond the $50 each cardholder is liable for, they have a strong incentive to identify - and immediately stop - fraudulent card use as early as possible in the fraud cycle. Otherwise, they face losses up to the remainder of the entire credit line on a credit card account. For that reason, banks use sophisticated neural net software to score the likelihood that each individual transaction might not be genuine (a gargantuan technological challenge made possible by innovations of the past decade) and thereby identify individual transactions that don't look like they were conducted by the accountholder - then they use Adeptra (or their manual calling operators) to contact the cardholder and verify these transactions, hoping to catch fraudulent card uses before all the money has been spent/lost.

NB
What is interactive alerting?

VK Think of it like a voice response unit that dialed you (or that emailed you, text messaged you, or contacted you in any number of ways)! When banks give us only phone numbers to use, it works like this: Adeptra rings a cardholder on behalf of the bank, establishes their identity by asking authentication question(s) and comparing the response to the expected result, then asks them to verify specific transactions (as many as five transactions are verified in a single interaction) - when cardholders verify the transactions are legitimate, our system then interacts directly with bank systems to update the cardholder's profile (so these transactions won't continue to trigger suspicion); when they can't verify the transactions, we connect the cardholder immediately with a fraud analyst at the bank who is authorized to take the appropriate action with the cardholder to protect the integrity of their account.

NB
What are the benefits of using interactive alerting?

VK In the fraud example above, the critical financial benefits are reduced cost to contact more cardholders sooner - and, by reaching them earlier the actual amount of fraud avoided is increased over the current method. Additionally, the fraud analysts' jobs are more interesting because they spend less time on rote "all legit" calls and more time dealing with the fraud cases. The bank also gets a more detailed audit trail of all attempts, complete and incomplete interactions, and resulting actions.

NB
How are banks using this technology for credit card payment collections?

VK Payment collections is another example where being able to reach and interact with cardholders more cost-effectively is of great value to the bank. The sooner and less expensively cardholders can be reminded to pay their amount due, the more "early delinquents" can be handled by our automated interactions while the bank's experienced workforce focuses on the later stage, more difficult collections challenges. ROI's in excess of 100 percent are being realized by our clients today using this service.

NB
Can you describe ways in which banks can get in touch with customers when they suspect fraud?

VK I think I described the most common method above, but our Solutions also have the ability to follow a complex contact methodology (we call it a "Find Me Follow Me strategy") that may involve numerous attempts over hours or days, more devices than just telephone such as digital text messages to mobile phones, e-mails, instant messages… basically any medium the customer chooses for the bank to contact them.

NB
How would a bank proceed with setting up interactive alerting?

VK Fortunately, we've built a platform that makes deploying turnkey applications for our clients a fairly straightforward process: clients spend one day with us choosing the parameters of Find Me Follow Me strategy, data file formats for input and result postings, script treatments and reporting options. Also covered are the various interaction modalities - touch tone, directed speech, natural language speech - and how they are to be deployed. For example, certain kinds of data capture can only be accomplished using speech recognition and others are more amenable to asking the user to enter certain keystrokes on their phones. We consult with clients to match the proper modality within each script, to maximize the effectiveness of their program. After we get their approval on the "application definition" we document from the initial workshop, in a matter of a few weeks the client is test-driving the completed solution before approving its deployment with cardholders.

NB
What else can be done to help banks react to suspected credit card fraud?

VK By integrating our service directly with the bank's card management system, we enable them to chase down these verifications in such a timely manner that some banks are reconsidering the paradigm they use for determining whom to call. Whereas (with the traditional people-based methods) it isn't necessarily cost-effective to call below a certain predictive transaction score (from the neural net software described above), with our lower-cost approach they can affordably contact a much broader range of suspicious transactions in order to catch fraudulent use sooner (and save more by avoiding higher fraud losses!). Also, speech technology is making the interaction so user-friendly that in many cases clients are considering using proactive outbound contacts in other applications as a way to preemptively decrease the flow of inbound calls for repeated inquiries or common questions. We are currently providing Just In Time Response Solutions to (in addition to our core credit card client base) a leading insurer as well as to a few airlines and travel companies. The insurer uses interactive alerts to provide an "opt-in sales presentation" that pre-qualifies interested prospects before they get in contact with a knowledgeable (and expensive!) sales rep in the call center. Airlines use our service to let their travelers know about schedule changes, gate changes, flight delays/cancellations, etc. in order to facilitate customer retention in the face of the inevitable glitches associated with traveling these days.

NB
What are your plans for the future?

VK We are continually innovating on several dimensions to make sure it remains a "no brainer" for our clients to use our service instead of building out this type of functionality on their own: we're improving the resiliency and scalability of our infrastructure to accommodate larger campaigns; we're enhancing the functionality we deploy in applications to improve the quality and ease of the customer contacts and the timeliness and accuracy of the data captured; finally, we keep improving the ease of use and performance of the reporting and integration engines we use to interface with client systems. The aim of all these innovations has been, of course, making it easier for us to deploy applications in new areas such as more complex marketing interactions like account activation or subscription renewal or address correction capture - guided always by our focus on understanding, operationally, how to provide high ROI by enhancing processes that clients are already aware can be improved to great financial benefit. Deploying exciting new technologies like speech recognition - and some day soon, even voice verification - is not in itself important; our goal is to do so in ways that enable us to deliver the business benefit to our clients.