Are you constrained by existing technology investments to compete with digital competitors?

In this 30 minute interactive webinar we will outline how Luna Connect and Inpute Technologies have implemented solutions to streamline lending processes, increase revenue through digital channels, improve borrower experience, and reduce overheads with a digital workforce.

Learn about:

• Building a Borrower experience through Digital Lending
• Streamline integration with legacy systems using RPA bots
• Transformation in Lending and Underwriting Management
• Emerging digital trends in loan (lending) origination

Webinar Transcript

Jennifer Rowland:  Good morning, everyone. Welcome to today's webinar, with INPUTE Technologies and Luna Connect. My name is Jenny Rowland, and I am part of the sales and marketing team at INPUTE Technologies.

The topic of today's webinar is "Your Guide to Automating Digital Lending with Robotic Process Automation." Before we begin, just to let everyone know that this webinar is being recorded, and it will be available on demand, later today.

You can also submit any questions you have throughout the presentation into the Q&A function, and there will be a live Q&A at the end of this 30‑minute webinar.

Speaking today is Aidan Mullin, and he is a chief operations director at INPUTE Technologies. Aidan will be discussing Robotic Process Automation, and how a digital workforce can streamline the digital lending experience.

Aidan will also be talking through specific use cases of RPA being used to seamlessly integrate with legacy systems.

Some of you may already know Fergal Geraghty. Fergal is the commercial director at Luna Connect. Today, Fergal will be discussing the future of digital lending and the challenges traditional lenders are facing today.

Without further ado, I'm going to hand you over to Fergal. Thank you again for joining today's webinar, and I hope you enjoy.

Fergal Geraghty:  Thank you, Jenny. Good morning, everybody. You're very welcome to this morning's webinar on automating digital lending with RPA. I hope everybody is OK in the lockdown. Hopefully, it will be over soon.

Just a couple of slides for myself on Luna Connect, and who we are, who we're working with, and why we may be relevant to your organization. Again, my name is Fergal Geraghty. I'm commercial director of Luna Connect.

Who is Luna Connect? We're a fintech that's in Galway. We're been in business for three years. We're working with financial institutions on that lending journey, so we've focused on lending, on lending journey for borrowers right through to how they apply online through IP into the back end to the underwriting and lending teams.

Then integrating that data, that loan data into the back office system. A lot of those systems are legacy systems, which we'll talk about later. We do primarily work with lending, but we also work with other financial institutions, like pensions and insurance as well, which have been very similar journeys, very similar apply AML checks, and so forth.

From a lending perspective, we focus on three key areas, the increase in the loan book which everybody wants, getting more loans out increases the loan book which is difficult at the moment in this COVID environment.

Increasing the borrower experience would be the second area. It would give that seamless. That's a good understanding of how a borrower would interact with you digitally. Then also the increasing productivity and effectiveness, primarily on the backend from the lending and underwriting team.

We're seeing a lot of data on the backend and spreadsheets. A lot of issues then, as we focused this morning, on that integration of getting data into backup backend systems and out of backend systems.

Looking at the site here, this won't be that by PowerPoint. I'm on the clock here this morning. We've only got 30 minutes. Looking on one or three slides, that's all I have. The site here, we're looking at this morning is everything. It's some different version.

There's a huge increase in challenger banks coming into the marketplace. A lot of them don't have that legacy back the office systems/ banking systems, they're very tech‑savvy. What about [inaudible 3:52] there, a lot of it's cloud‑based, application's very agile. The revolute to the world, the advanced [inaudible 3:58] of the world are coming into the marketplace.

Lending is still with a traditional lender. It's still with the institutions. Here in Ireland, the four federal banks still do a lot of the consumer‑based lending, but that is changing. The non‑banks and it is changing over especially in SME. SME is a big change. It's happening where a lot of the traditional lenders are giving way to the non‑banks.

Then looking at some of the finance of leasing and credit unions delivering 80 percent abandoning offline. It's not completely offline, but it's a hybrid of online and offline where maybe an application online comes into an inbox. That inbox maybe happened to [inaudible 4:43] it on a spreadsheet or maybe [inaudible 4:44] a backend system and so forth.

All other banks, they're removing all that manual‑based interactions or silos of data and so forth. Looking at distilling that problem down into maybe four key areas, and this is what we're seeing in the marketplace. As I said, increased competition, digital savvy.

How do I up my game? How do I become more competitive and attract those maybe younger members who are used to a mobile forest or the revolutes and so forth. It's all about quickness. It's about delivering products to market very quickly.

As I said, the competition is everywhere. Amazon and the US are delivering SME loans into the marketplace and [inaudible 5:27] to Europe. PayPal [inaudible 5:29] working capital.

Even in Ireland. If you take the Harvey Norman at [inaudible 5:34] , they're working with Flexify on zero‑based APRs for purchases of their wares. In other areas is that interaction of that long journey.

That face‑to‑face, the paper‑based applications, the bank statements, the payslips, the emails, the phone calls. A lot of work of just getting to it to a loan application and then delivering that loan data to the underwriting and lending teams. That does slowdown that whole process.

We try and quicken that up and efficiencies around that. Then the underwriting teams completing out that data and getting it into the backend systems. That brings us to the four‑point constraint for existing technology.

This is not just a lending environment. Anyone of the team members here would have a huge experience of working with a lot of industries not just financial services, of technology footprints that are there for [laughs] 40, 50 years. I know the banks of 40, 50‑year‑old technology, and they use our RPA too, to extract data and import data into their systems as well.

Moving on to the last slide, a brief overview of Luna Connect. Who are we from a platform perspective? Luna Connect is a platform. We don't build bespoke products or bespoke platforms for lenders. It's [inaudible 7:02] platform. It's very privatized. We have a very detailed road map that we're working towards.

At the moment, we focus on the journey from that acquisition [inaudible 7:14] through...

[notification tone]

Fergal:  ...getting that, from a lending perspective, that loan into the backend system. From acquisition that's very much focused on working with marketing teams. How are you acquiring the borrowers? What are you doing for digital capacity? What are you doing offline on helping the lenders then bring attractive borrowers, bring them on to the site?

From an application point of view, it's very much focused on a seamless user interface. Remove as many frustrations as we can. We've all used applications online where we've got very frustrated, but also brilliant technologies like [inaudible 7:57] .

Our next webinar is on open banking with [inaudible 8:02] , so that should be, hopefully, of interest to people. Then onboarding, everybody's financial perspective is regulated. Sorry, not everybody's regulated, but if you're regulated...I'm not regulated. I have to adhere to the KYC amount checks for the central bank so it has to be seamless.

Again, I've seen KYC‑AML that brings you to a mobile. If you're applying online, you don't get verification there. Then bringing all that data to the backend of processing. To the underwriting and lending teams, to help them process quickly and also integrate with the backend systems of using RPA. At that point, maybe you're bringing data onto the platform.

Then analytics, Luna Connect is very much a data‑driven company. Our bread and butter is data. We're all about looking at the data that's coming in our front door from a loan lending perspective. How many loans have come in today, this week, this month? Are we having KPIs? How can I analyze that data?

Again, using RPA, or if the APIs are available, bringing data from backend systems. Bringing that loan data in. We have the ability to apply machine learning to loan data, to understand maybe the history of the data, or look at some of the statistics against lending.

Then that brings us to the focus of the webinar this morning. Integration. As I said, Luna Connect is very much an open platform, open API, where we welcome working with other platforms that are installed in lending institutions.

We integrate with a whole host of platforms from open banking, to KYC, to credit checks, to any type of application that's used on the journey. If the APIs are not available, we can then use the product, RPA, which is again not to steal [inaudible 10:02] . It is simple to implement.

A lot of it comes out of the box. It's very much configuring the workflows to extract the data. I started to input the data and then the ability to extract the data. It can work in the background from the way in the [inaudible 10:22] on the tender part.

On that note, I hope I've given everybody a good overview of who we are, our platform. With a lot of content on our website on the main webinars to go into the product. Some of our customers, they talk about the value that they're driving from our platform, and how they're competing with our competitors in their space.

Please post any questions on the Q&A. We will have a Q&A session later on. Our email is at the end of it as well. Under that note, I'll hand it over to Aidan.

[pause]

Aidan Mullin:  Thanks, Fergal. Good morning everyone. As Jenny said in the introduction, my name is Aidan Mullin. I'm the operations director of MP Technologies.

At MP, we've been automating business processes for organizations for over 20 years now. We have the last 3 to 4 years in delivering what we refer to as middle layer solutions using RPA robots, or as sometimes referred to as the virtual workforce.

Today we're going to speak about how you can implement a virtual workforce to help your human workforce to run your organization more efficiently.

I'm going to, first of all, introduce RPA. What RPA can do and what it brings to the table. Then we're going to go through some practical use cases that we've implemented. How we've implemented solutions that delivered real‑life and solutions to customers.

What is RPA, first of all, and what can it do? RPA is pretty much software that runs on a computer. That mimics the activities and actions that the user takes when they're in technical computers. Clicking around, moving the mouse, entering data. All these things that a computer does, that's what the software is doing.

It can interact and operate with any application, and it can take actions that are specifically laid out. That's for Robert's take. A good use case is where you've got repetitive tasks that you want to be completed. The robot can then take them as long as there's a very structured fit around that.

What does RPA bring? Its key benefits where you've got a task that is repetitive and has a number of steps that need to be completed. You can find when a human is undertaking this, they can find it quite boring, monotonous. They can switch off and begin to multitask or think about other things.

When that happens is you get mistakes that were made from manual data entry. Things have to be repeatedly done, and it's not rewarding work for employees to complete. That's what a robot is.

It provides better accuracy. It can complete these things. It doesn't get bored. It completes these tasks exactly as they've been programmed to do. It can take data from a source and data enter it into a screen a million times and will never get fed up doing it.

What it allows to do is take tasks that have these multiple tasks to be completed. Where there's fixed rules that it will apply that doesn't require any thinking about variables that are in the gray, and how they might impact the process.

If there are these variables, then the task gets moved. Allocated to a human for them to be able to interact with the case. The majority of the cases that don't fall into this variable situation, they can proceed and be processed by the robot. The robot will allocate the others to humans. Humans can then only interact with the cases that require their intelligence in order for them to be able to complete.

[pause]

Aidan:  How have we implemented this before? The first use case we're going to talk about is one of our customers, An Post. An Post implemented RPA as part of their Web portal customer onboarding. They had a Web portal they were deploying that they wanted to bring all of their customers on board.

Historically, An Post have sold many products over the years right down to the government search and bonds of their agents for the NTMA, to prize bonds that people might be familiar with, right down to when you opened your first post office account as a child.

When they were on board and customers on this portal, all of these accounts need to be linked with particular customers. They did a migration, and the migration was somewhat successful. It was understood there was quite a lot of information that wasn't transferred across to customer accounts.

When the customers began to interact with the customers, the customer might highlight that they believed that there were other accounts that hadn't been correctly allocated against their account. In this case, the customer service agent would initiate a customer account trace request.

What they typically, normally have to do is they go back through the legacy An Post systems. [inaudible 15:42] had many of them for all the different products and services that had been implemented over the years. They troll to them to try and find the accounts that's related to this customer.

When they complete this task, they would then come back and find the appropriate accounts [inaudible 16:01] correctly and associate them with the accounts. What they look to do is implement an RPA solution for this.

The robot would receive the request for a customer account trace. It would then go and view all seamlessly and integrate with all of the legacy systems. Using the things like fuzzy searches or fuzzy matching, it would use data such as the state saving customer number, the customer's name, customer address, date of birth, old addresses.

It would merge this data and leverage it within all the legacy systems to retrieve back results that might be relevant. Accounts that might be relevant to this particular customer.

It would allow us to bring back those details and verify that these are probably accounts that were linked. These were less likely to be linked to the customer and these were unlikely, and present these back to a caseworker.

To deliver that relevant data back to the caseworker so that they could then make intelligent decisions as to whether accounts should be included or not, or not included if that's the case. They will then, within the customer care system, link those accounts which would then appear on the customer portal.

This is a good example of where the robot was implemented to take out all the mundane legacy data trolling, data mining that needs to be done, present them back up the relevant information to the caseworker so they could review the relevant information and make an intelligent decision.

AXA Insurance, we've also partnered with them in implementing their RPA strategy. They had a use case where they receive paper mortgage statements from which they need to capture the payment line items and use that data in an employee benefit scheme, if they have. They had members of the finance team, so high value people who were working with these statements.

They were capturing the data from these statements, and manually entering the data into a spreadsheet, and then calculating the allowance. It was all very manual, and it was a lot of data capture of mundane data work that people were doing, and it was error prone.

They found the manual capture and the manual calculation was giving different calculations coming up, because there were different interpretations, and a lot of data entry errors. The main reason was because these high value workers found it very frustrating, working with this low value work that was giving low value to the business, and it was taking a lot of time.

Again, we looked at implementing an RPA solution. First of all, we automated the data capture. We linked it in with an OCR capture server, so that when the documentation came in, it was automatically processed. The line items were automatically extracted. We could then eliminate keying errors by automating that data capture, and validating the data.

By having data validation that the robot implemented, it meant that if there was any question mark over any data, we sent it over to the user for the human to interact and say, "Yeah." For example, all the line items don't total to the statement amount, we could bring that to the user.

The user could then make an intelligent conversation and say, "Oh, there's lines missing here," or, "There's lines that were read maybe incorrectly." Whatever the action is, they can quickly and easily take that data.

But they've only been asked to interact with a small minority of interactions. There's not that mundane difficulty or error, due to mundane tasks. Then once the data's been verified, the robot has all the information that it requires. It has the data they need automatically extracted, along with the manually verified data.

They can then verify the claim that's required, and integrate directly with the legacy payment system, to input that payment coming out of the process, into the system. That, again, was a good use case where the human is still used from the human is good at and the robot manages the steps that it can manage, within the parameters in which it can operate.

Another practical use case is our integration with Luna Connect. The Luna Connect digital lending platform, the data is captured during the loan application process, and needs to be transferred to the back end legacy loan management software.

Although the Luna software itself is a fully‑featured modern API, to retrieve data from it, the legacy systems didn't provide any program interface that allows us to send data into the database.

The only way that we could capture the data is via the front end application screen, so basically, going in the front door of interacting applications. For a user to manually key in the data was repetitive, again, and it was mundane. They were keying data that had been already captured previously.

It seems that it was overkill to have to go and capture that into a second system again. But this information and the captured information into the loan application management software was a key task.

Our RPA solution, again, we had seamless integration between the robot and Luna. We could automate the human task, that was the data entry, and the repetitive data entry task. We made sure that we capture all the relevant data that from the application that's come from Luna, we can gather all that information that's important and is needed to manage the loan application.

We can implement that into our back end legacy system, faster enter end processing, and we increased the data quality, as well. That was the third use case.

Other business cases, there's loads of them. There's many of them out there, where RPA, either independently or linking it in with something like an OCR engine, can be used to implement an automated process. There are many use cases across many industries, that can be used.

How do we begin your RPA journey? What we suggest is that there are three steps. We look at a proof of concept. We pick an initial number of a particular use case, and we would implement a PRC.

We typically do that in one to two weeks, so it allows everybody to be able to look at the systems we're going to be integrating with the processes that are involved, the steps that are involved, and the benefits that are going to be returned from that.

We then run a pilot for about two to four weeks, on approval of the proof of concept. We develop additional automated process that encrypt and implement that, so that it can be seen as our first initial robot, and the benefits that can be reaped from doing that.

Then, we rolled that out. Depends on the scope, you either have a full rollout for many different processes, or we implement specific use cases that are required.

[inaudible 23:25] the slides and the use cases that I wanted to use today to try and give you some information on how RPA can be implemented to deliver automated processes, specifically around integration with legacy systems. I can see there's some questions that I'm going to hand back over to Jenny, and we'll move into the Q&A.

[background sounds only]

Jennifer:  Perfect, thanks a million, Aidan. We're going to move into the Q&A section. There's a couple of questions here. The first one is for Aidan, "Where does the bot run, and what's the difference between an unattended and an attended bot?"

Aidan:  The robot runs on the infrastructure. You mentioned there, in the question, about attended and unattended. An attended robot is a robot that runs on a user's machine. It requires a workstation, and the workstation's required to be dedicated to that process while the robot's running. A user interacts with that, and it's running on a client workstation.

An unattended robot is the next level up. That's where you can create the robot to run as a background process, maybe in a backend server, where you can have that running in the background. It will interact without any user interaction. It runs as a, what you've referred to the presentation as one of your virtual work force, so undertaking tasks without any user interacting it.

They can reside either on the backend server, or on client workstations. The difference with the client workstation, with the attended robot, it does require the user's machine not to be used for another purpose. Whereas, a backend, unattended robot, it runs as another shared service on the backend server.

Jennifer:  The second question here is for Fergal. It's, "How does Luna Connect integrate with the bot or RPA?"

Fergal:  Jenny, thanks very much. Luna Connect does, as Aidan said there, as full APIs, and do RPA affect with the bot as an API? When the data's complete on Luna Connect, as I said before, whether that's a loan, or a pension, or insurance policy, that gets completed through the process of Luna Connect according to the business rules. Then the bot would be polling.

It's an unattended, it would be polling Luna for the completed status. If it's attended, we can click something on Luna Connect, so it then transfers the data to the bot. The bot takes over then, and puts the data into the backend.

In reverse, we can integrate with the bot to pull specific data out of the application of the back office system for pulling that into Luna Connect.