To watch the recording in full, click here, otherwise, you can read the transcript below.
Lenora Thomas (LT): Hi everyone, my name is Lenora Thomas, Director of Customer Success at Trust Science.
Thanks for joining us today in our webinar, “Automated Lending with Vergent and Trust Science”
We're excited to give you the first look at the Trust Science and Vergent integration. Now, with Trust Science, lenders can score credit invisibles from within their Vergent LMS. Trust Science offers the first step in automated decisions. We use machine learning and AI to give you predictable scoring so you can increase originations, reduce defaults and eliminate errors in human judgment.
We know credit decisioning can be challenging. There are many manual steps, multiple systems and the right data is missing.
Today, you'll see how easily you can use Trust Sciences functionality from within your Vergent LMS system.
You can eliminate manual work, centralize information and gain access to the right data.
So to kick us off, I'd like to extend a special welcome to our panelists.
We have Herb Hulme, Lender Relations Specialist with Vergent LMS, and Brian Katis, Chief Product Officer here at Trust Science.
Herb will demo how from within your Vergent account you can access Trust Science features like automatic credit scoring then Brian will demo additional benefits like how you can gain consent from borrowers for social and banking data.
The demo portion will last for about 30 minutes and then we'll open up the webinar for a 15 minute live Q&A.
So let's get started over. Herb, walk us through how you can use Trust Science from within your Vergent LMS.
Herb Hulme (HH): Good morning, everybody. Inside, you know once you guys have done your subscription with Trust Science and agreed upon what your minimum credit scores will be and your maximum credit scores will be they will send you a set of credentials, which you will send to us along with the credit scores and we will set this up in the back of our Vergent systems.
And the main thing on the screen that you want to notice is that there's a minimum of a 500 for minimum credit score and a maximum of 600.
The minimum amount of information we need from a customer, of course, is customer name customer address, including city, state, zip customer valid phone number, email address.
If this information is not in Vergent and you tried to do a Trust Science pull, you'll get back a falling error message. Go fix this. And then you can do another pull.
First, let's start out and saying that there's three ways that you can get messages back from your Vergent software. And that is if somebody falls before below the 500 credit score, you're going to get back a red light, which is going to be rejected. If I loan credit score falls to clean the five and 600 amount you're going to get a review process that's going to go to review manager. In this case we've set up Betty Wilson as that approval manager.
And then if it's above 600 you're going to get a green light. And this loan will be approved for processing. So all you have to do is pick your low model, just like you do today and hit the Trust Science scoring checkbox and hit next.
And this particular one, we're taking a loan out on Randy Wallace, he's going to get back a score of 574 that is being sent to us from Trust Science, in this case, we will send this out to review with our approval manager.
The proof manager screen shows up and they have the right to either approve or reject this loan in this particular situation. Betty Wilson is putting in a $200 maximum credit limit on this person.
And what you see back on your general information screen in your Vergent software for Randy Wallace is a score of the 574 has been approved by Betty Wilson on a particular date at a particular time. And of course, we timestamp all of these transactions. So we can go back in history and check this out.
Then you soon as you get that approval, you can be able to process your loan, just like you do currently in Vergent. Next, we're going to do Ashley. Ashley is wanting alone, we're gonna go through Trust Science for the scoring. She comes back with a review 550 which falls into the review category. We're gonna send this out to automatically to Betty Wilson for review.
Betty Wilson gets the review on this particular customer. And in this case, Betty rejects this by maybe looking at some other documents or seeing that the information does not match up. So at this point in time, we know that Betty Wilson has rejected this particular loan for Ashley.
And then Alicia Coleman. We're going to do a loan for her. And she falls below the 500 scoring limit. So we are going to give her an automatic rejection. And that shows up again under the general information that this alone has been rejected.
Dominic is applying for a loan. So we're gonna do the same thing for him. And in turn, he is above the 600 so he gets that immediate approval and you'll go right into the loan processing stage for this gentleman.
That's our side. Very, very simple, very straightforward. When you're interested in doing a Vergent and Trust Science integration give either Jennifer or myself a call or an email. We can get you out of, quote, we can set this up and have you running. Thank you very much. Back to you Lenora.
LT: Thanks for the great introduction Herb. We really appreciate it. We’ll now switch over to Brian to see how lenders get access to additional Trust Science features like the ability to gain consent from borrowers for social and banking data.
And we'll have him walk us through how lenders can eliminate manual work and get the right data by the borrower.
Bryan Katis (BK): Great. Thanks, Lenora. So I'm in the Trust Science lender portal right now and starting where Herb left off with Dominic
Dominic had a 602 so this is past that 600 threshold, as you saw. And what I'm going to show you now is if any event that we wanted to get some additional data from the consumer what the steps would be to do this. So to do this, it's a very simple flag that would be set in Vergent to say send a mobile invite. And then an invite will be sent via SMS or via email to that borrower. So in this case, it was sent via email.
And Dominic receives an email that I, the lender. In this example, Quick Cash would like to have them use Trust Start. Trust Start is our web as well as mobile application that enables us to collect that consented data both social and banking data from the consumer.
So I will then click on connect to Trust Start. The borrower has been taken to our web application where they're showing the terms and conditions.
They're also showing the frequently asked questions. What will they to be used for, why do you need my data, so on and so forth. And once they agree to these terms they can accept and then the borrower is presented with the details to reconfirm that they've applied for this loan that this is the amount their contact information. And if they agree to this they can confirm.
They’re then presented with our instant banking verification solution we provide access to over 15,000 financial institutions from banks to credit unions.
And in this example I will click on Trust Science bank in the top left corner.
And then the borrower is presented with what the data will be actually used for and collected. So we will be accessing the account transaction history. Their account number, as well as their name, email address and phone number.
I'll agree and continue
I then add the credentials for the borrower their username, their password.
It then takes a few seconds to a few minutes to authorize the bank and then once this is done the bank account has then been linked to the loan application and all that history is available for analysis purposes. And I'll talk more about that in a few minutes. Next, optionally the borrower can provide consented social data. So this is an area where Trust Science has a number of patents in this area.
And we've been able to see with studies like a Duke University that social data is as predictive, if not, more predictive than the traditional financial data that the credit bureaus have been using for many years. So in this example, I will click on LinkedIn.
And we're using the standard LinkedIn API to be able to authenticate I'll copy and my password.
Click on sign in.
And now I've linked my LinkedIn account with my profile as well as my network as well to the loan application and this will then enable Trust Science to create a social profile. And that borrower as well as their network using that consented data that the borrower has provided. I can add additional social media sources or I'll click on Next.
And then I have the option to download our mobile application. The mobile application will provide additional consented data directly from the mobile device.
So I will skip this at this point.
And the very last step here is the borrower could create an account with Trust Science. By doing this would enable them to view the status of loan processing, as well as be able to reload and other features.
I will skip this and go back the score has now been provided. And you can see that the score has been lowered and you can see it's lowered for a number of reasons. One, that the monthly payroll deposits, do not match the state of monthly income that was stated on the loan application.
There have been several short term loans recently drawn down from another short term lender, the name that was provided on the loan application does not match the bank account.
And we also see that there have been no payroll deposits found in the past 90 days. So these are all flags that we've been able to find in our proprietary bank categorization model.
So if I scroll down, one of the things that we provide for each of the borrowers, then that provides banking data is first to summarize.
So we can see that there have been 12 payday loans in the past 90 days, and there has been zero payrolls added and then none in the last 90 days as well. And then if I scroll down to the detail I can also see that this isn't Dominic. This is actually John Doe as the account holder.
I can sit their current balances $20 and then I can also see some additional transaction data, some of the different debits and credits that have happened over the last period of time, and some different repayments that they've made. And so based on this additional information, it helps the lender determine is this someone that they should be extending a loan to.
So with that, I will hand it back over to Lenora.
LT: Thanks for taking us through Brian, it's helpful to see how Trust Science enables access to the right data for each for over so lenders can speed up their decision. And with that wraps up our demo with the walkthrough of the trust sign solution with our predictable scoring, you can increase originations reduce defaults and eliminate errors in human judgment.
Now I'd like to open up the Q&A portion to our audience. And we have a couple of questions it looks like it. Okay, so the first question is, do you have any customers live with this integration yet?
BK: So this Bryan, I can tell you that. Um, so we do have this integration with a lender that is up and running at this point. They started with the short term pay lender use case. They will then be expanding into installment and title in the next phases and so as Herb mentioned in the demo, when you choose the type like cash advance whatever that might be you have then that radio button or the checkbox to say use the Trust Science scoring. And so right now they're selectively choosing that and then once they bring all their loan types onboard.
Then we would automatically do this for all their various loan types. So we would support any of the various loan types that are available out there.
Whether it be automotive, whether it be payday, installment or title. Those are all supported. So if you're interested. We definitely would love to hear from you.
LT: Excellent. Okay, we have another question here. Who can view the Trust Science Information with the banking information?
BK: So I can talk to that as well. So that information, the one I showed you was in our lender portal and so that is role-based security. So it is specific to the person that is assigned as the financial consultant, the loan consultant. That is the person that's assigned it would also be assigned to the manager. If they end in the hierarchy so that they would be able to see that. But it's restricted to only the people that are required to see that information herb.
LT: Herb, do you want to talk a little bit about how that would be restricted inside Vergent?
HH: Well, as usual, we set up customer loan models inside of Vergent and we always are putting mins and maxes and different controls on how loans to be presented to the cashier so that they stay within the bounds of their states that they're located in. So a lot of controls, go back and forth between ourselves and Trust Science to keep everybody compliant and keep everybody running forward.
LT: Excellent. Right. So the next question that is asked.
I'm just going to put this into a bit different standpoint: the Trust Science score will pull a traditional credit score upon initial pull from within the program, then the customer must opt into banking and social components.
Brian, do you want to take that one?
BK: Sure, so we can definitely leverage a traditional Bureau score. So for those lenders that are using a traditional Bureau score including the attributes we can leverage that by no means is it required the traditional Bureau score only looks at roughly 3% of the data that's financial and structured in nature, while we look at that. We also look at unstructured and non-financial data. So we look at the other 97% of the data that's also available.
On those borrowers, we look at a much more comprehensive view of the data. And so we can look at whatever data is being provided by the lender via the loan application by Bureau data whatever else there might be, ID, fraud, bureau, whatever else we can leverage all that and then we complement it with. Our the consented data as I showed you, which could be the instant banking verification can be social media. We also do a public search of the web or will find social will find court records and other information that's useful. And then we also have our own proprietary data sources that we use as well. So this provides a much more comprehensive view of that borrower.
And I'll also note that we're fully FCRA compliant. As well as DPPA, GLBA. And all the various acronyms that are out there as well as FCR compliant in Canada. And so we ensure that we're following all the various regulations around this.
LT: Okay, that's perfect.
So Herb, Brian either of you can answer this one.
What sort of timeline can a user expect in terms of the setup time for Trust Science and Vergent?
HH: Brian. Start with your side.
Sure, sure. The advantage of us is our productizing in this integration is the lion's share of the work has been done and it's just really configuration as Herb showed. And so what you would do is working with Trust Science, being able to we would work with you to build a customized model that factors into your risk tolerances, the type of loans and products that you leverage this isn't a one size fits all like a traditional Bureau score. Whether I'm getting an auto loan or applying for a credit card and I'm getting one score. We actually tailor the model to the actual loan terms.
And again, the risk tolerances of that lender. Some are more conservative. Some are less conservative and so we work with you to optimize decreasing the defaults, increasing the originations as well as helping you risk price the solution. And so it will vary based on the customer. We typically target six weeks for implementation.
HH: With Vergent it’s very straightforward. Very simple, just like we do all of our other integrations is customer, please get in touch with either Jennifer or myself either by phone call or send us an email that you want this integration with Trust Science. We will, in turn, get you out of, quote, once it signed and dated and comes back in.
We're going to get you set up probably within a two or three day time period. Less than a week would probably be maximum, and that'll be just working on your software, just with a few clicks and setup information. Once we receive your credentials and your credit scores.
LT: Excellent. OK, the next question is, do we offer an API to hit Trust Science from their solution.
BK: So Trust Science offers API's that enable a lender to send a scoring request to Trust Science. And again, that includes any of the data that would be available from a loan application to Bureau to whatever it might be. Trust Science provides a second API that sends that scoring report back and then third, we provide the ability to for the lender to send us the performance data on a daily, weekly, monthly, whatever interval. Makes sense so we can continuously monitor what's happening with the performance of the loans. So we publish those make them available and Vergent has leveraged those API's in order to productize this integration. For example if you are working with another, if you have a homegrown system or if you're working with another provider that you could use these APIs.
LT: Excellent. Okay, so next question up is how have customers successfully incorporated Trust Science scoring into their process.
BK: First, so I can take that. So we've had one customer as an example where they were able to decrease their defaults by more than 6% we had another customer that added Trust Science and were able to increase their loans to a new group of users that they traditionally would not include, so we're able to find additional borrowers. We say sifting the prime amount of subprime that are really good borrowers that were just traditionally invisible.
And I know based on the criteria that was that was being used. So this really opened up the ability to increase those originations bye adding this new group from an automation standpoint, we've worked with a customer that was able to bring all their lending qualifiers into Trust Science, they did not have to manually review the banking data or the Bureau data and this saved a significant amount of time.
But on average, we see things where it could be 10% lift in our ROC. We've seen it in the 20% are OC lift as well. We've had some banking customers that said, hey, if you can show me three quarters of percent of left. You know, I'll sign up for that every day. But in general we see double digit lift.
LT: Okay. Now this next question. I'm not sure if it would be for Herb or for Brian
Is this software integrated with state required payday loan databases such as Veritec?
HH: Well, probably, that's mine. I'll take that one. Vergent has always been hooked up to Veritec databases and every state that they are in. If you go out to their website, you will see that we have been a Class A integrated with them. Ever since they started their first database. So that will always take place you want to be in this business. You've got to be hooked up to state databases as required.
LT: Thank you so much.
OK, the next question up is, and someone answered this, Brian is.
How is this any different to using traditional credit scores?
BK: Sure. So again, the difference between the traditional credit scores are they usually are only looking at 3% or so maybe up to 5% of the structured financial data that's available over the years, they've been trying to add additional things like utilities and rent and some cell phone bills.But again, that's still that structured financial data. What we do is we again we will leverage that it's not that it's useless. We will leverage it but then we will augment it with the data that we are able to get. And again, we get this from our proprietary sources from the consented data as well as public sources to be able to get a more comprehensive view. So we're not just looking at maybe 3040 indicators, we're looking at thousands and thousands of indicators. The other thing is we're using AI and machine learning. And so this is where we're able to look at the data, find patterns correlations of the data were looking at one attribute one variable one indicator by itself might not be indicative but looking at a series of variables on in in total provide a lot more insights and so we look at and not every borrower provides the same amount of information. We do do this search in real time.
To find what information is available I and again we'll use the traditional sources. But then we're also using alternative data sources and other sources that provide a much more comprehensive and much more realistic view of scoring borrower, especially when you're talking about the subprime market.
The traditional bureaus are very strong in the prime market, we're really focused on the subprime lender, because that's really where they're invisible, or they have the thin the no file and and they all look the same. And you really can't differentiate them using the traditional bureaus.
LT: Excellent. And Herb what about for your customers who are using Vergent right now. Is it dramatically different to Trust Science scores versus the traditional credit scores and Vergent.
HH: I would think it'd be much easier because you've got such a simple path to get there. But the big thing is now you're pulling up a credit report from somebody with some of the vendors that are out there where this will be automatic for your case years or least go through a review process by your approval manager and cut out the need for a cashier to understand a credit report and make some type of logical decision.
LT: Okay. That's perfect. Alright, so the next question up is a two parter.
So the next question is, If customers are asked to provide their banking and their social, are they required to give approval and consent for that data, which is important.
And the second is our is Trust Science able to provide a score without doing the banking and the social?
BK: Yes, so it's completely optional. And yes, we can provide a score with just for example, loan application. And then the public sources and the proprietary sources that Trust Science brings to the table.
LT: And an additional question for you, Brian.
How does Trust Science come up with the cut off so they approve the decline in the review?
BK: Sure. So what we do is we tailor this again toward directly towards the lender and so we have what we call our scoring tool and we will provide this to a customer based on historical analysis of their data of their business. What they're aspiring to to do with Trust Science and we will then provide an upper cut off and a lower cut off. And as you saw with and use example, there was a 600 at the top of 500 at the bottom.
That will vary by lender. In some cases it might be 640 and 520 we will work with the lender on this and we actually provide this tool. So that the lender can do a lot of modeling and simulation where they can just simply change those numbers and then it will show the impact on default rates on the originations on profitability and then just tweak it to find their sweet spot of what they're trying to achieve and what we see over time is that that manual review window which in heard example was 500 to 600 100 point spread that tends to get reduced over time. Once the Trust Score goes into production.
LT: Okay. I think that is all the questions that we have time for today. So that wraps up our webinar.
If you'd like a personalized demo and how we can help you automate your decision making process and add data to score your thin file or credit invisibles please visit us at trustscience.com/learnmore or email firstname.lastname@example.org.
Thanks Herb and Brian today and thanks to all of you for joining us. Enjoy the rest of your day.