blog - Modelshop - Page 2
az

No Code Mortgage Loan Decision Application

At a time when automation and big data are at an all time high, the speed with which you can manipulate and make decisions on data can be the difference between life and death for a business. At Modelshop, we provide businesses of all sizes with the tools they need to make automation and data decisioning as frictionless and easy as possible all within a no code environment. Our clients have the ability to not only analyze their data with our easy to use user interfaces, but also modify their data and data environment with ease and immediately see any changes flow through to the rest of the application in real time. 

To show how easy it is for Modelshop to automate your data decision making processes, we’ve put together a short video showing the steps required to build a simple, effective, and easy to use decision application for underwriting mortgage loan applications. Simple applications like this one can take less than a day for a Modelshop analyst to build, but allow users to reap the significant benefits that come with access to fully automated applications with fully integrated machine learning and REST APIs.

This demo is meant to show you some of the computational capabilities of Modelshop, while demonstrating how quick and easy it is to build a model from scratch. These capabilities can easily be adapted to solve whatever problem your business needs solved. Whether you need a simple credit decision making model or a large scale marketing tool, Modelshop has the best tools available to automate your decision making processes.

Hello World AI Risk Demo

Modelshop is the fastest way to automate custom AI risk decisions across the customer lifecycle. Join me as I create an end-to-end “Hello World” style credit decision engine from scratch in five minutes.

Hello World Risk Decision Model

In just five minutes I will demonstrate a very simple, but realistic risk decisioning model, including deploying that model as a real-time origination experience and deploying screens that manage model performance.

No code credit AI

The no code software movement is real, and it is accelerating. After decades of false starts, we’re finally delivering on the promise of creating applications without code.

Most no code tools focus on building websites or simple business workflows, but they struggle with computationally intensive applications. Modelshop is different. It has been built from the ground up to deliver advanced analytic applications without code.

Watch the platform in action with this demo of a credit decision application built without code in less than a day.

Customer Churn Model – Part 2

This is part two of the multi part series where we will build a Modelshop model from scratch.  You can find the first part, here. In this video, we will build on the exploratory analysis we performed last time and use machine learning to determine the important features, weights, and cut offs for optimally predicting churn. We will also learn how to deploy that ML model and extract predictions along with confidence levels. Best of all we won’t have to write any code!

Customer Churn Model – Part 1

This is part one of the multi part series where we will build a Modelshop model from scratch.  We are making this series because we often get questions about what goes into building a Modelshop model…. What types of problems can be solved, how you get started and what the models can ultimately do for you.  The answer in short is Modelshop can be used to solve a wide array of business problems and is perfect for automating processes ranging from simple to complex. 

In this example we will be playing the part of an analyst at a retail brokerage firm.  The firm has seen significant customer attrition and went through the process of determining why.  With that research done, we have been tasked with doing the ongoing analysis to analyze the current customer set on an ongoing basis and identify customers that may be at risk of churning.  This way we can take proactive measures to intervene and stop before they close their accounts.  This demo will be broken up into several parts.  In this segment we will import our data and do exploratory analysis and test some hypothesis.  In subsequent videos we will implement a machine learning model do make automated churn predictions as well as create an application to share this information other stake holders. 

Can’t wait to see more? Check out part 2, here.

Payroll Protection Program Forgiveness Demo

Lenders are scrambling to implement processes to handle forgiveness applications for PPP loans. Using Modelshop to create a low-code, model driven application is a way to accelerate the process for lenders who might otherwise manage some or all of these applications manually.

In this demo, we walk you through an example of a friction-less PPP forgiveness application and then goes behind the scenes to show you how model-driven applications like this can be created without custom code.

A compelling feature of this model is the use of an analytic plug-in from a partner company Kodexa. Using Kodexa’s unstructured document analysis tools, we’re able to extract utility payments from uploaded bank statements and present them back for confirmation.

The ability to deliver a more intelligent user experience like this is one of the powerful benefits of backing applications with a low-code decision models.

Case Study – CSG Advisors

Client demands are growing for management consultants. Increasingly, firms are expected to be more than service providers, they are being asked to be partners in their clients’ digital transformation (CB Insights).

With over 30 years of experience working in consulting for the municipal bonds industry, Mark Kaveny has experienced this first-hand. To meet the growing needs of his clients, his firm is now developing custom applications that bring new value to their client’s processes.

Mark is part of a 7-person team at CSG Advisors that also provides advisory services to municipal bond issuers to finance affordable housing programs. CSG has traditionally served as a financial advisor, also providing specialized quantitative modeling and program advice.

Housing Finance Agencies are required to provide financial reporting to lawyers, rating agencies, investors, and the IRS.

CSG’s clients work in a highly regulated industry. Housing Finance Agencies (HFAs) are required to provide financial reporting to lawyers, rating agencies, investors, and the IRS. The regulations and accompanying reports are data and computationally intensive. Additionally, there are specific instructions on how data should to be entered, processed, calculated, and submitted to their tax council and other interested parties.

In some instances, electronic records had to be manually inserted into spreadsheets and required manual reformatting and data cleaning.

As an example, to build their arbitrage rebate tax reports every year, one such HFA client was manually downloading custodian bank web portal reports. They would put the hard copy of the report next to their keyboard and manually rekey tens of thousands of dates and dollar amounts into a spreadsheet. This was an extreme case, but even in some instances electronic records had to be manually inserted into spreadsheets and required manual reformatting and data cleaning. This process was awkward and time-intensive. The resulting spreadsheets were error-prone and required multiple review cycles. Once the errors were removed and the results manually verified, the spreadsheet summaries were manually copied into a document format and sent to the IRS.

There had to be a better way to prepare and transform data.

Mark saw an opportunity to improve this process. He knew there were better tools to prepare and transform data. After evaluating data and analytic platforms he chose Modelshop because of its configurability and ease of use. Modelshop gave his client a familiar tabular view but with a much easier way to add new data periodically. More importantly, Modelshop was an online tool that allowed Mark to collaborate with his client much more easily. Using their existing spreadsheets as a guide, he configured their model in Modelshop and showed the client his proof of concept solution.

Once the model was created, the new process for the client became quite simple.

Even as a new user, it took only a few days to configure the model. Once the model was created, the new process for the client became quite simple and straightforward. In the future, an employee will download data quarterly from their ledger system and custodian bank web portal and upload it through a Modelshop data connector. The model will automatically map the new data to each tax plan and flag any data quality issues. If new accounts are detected, they will be displayed in a dashboard and will enable the client to quickly and easily create new tax plans as needed and map the account(s) to the correct tax plans. Once all data relationships and data exceptions are resolved, a pro forma arbitrage rebate report is automatically generated that can be sent directly to tax counsel and the IRS.

What used to take 70 hours every year, now takes 2 hours, which frees up 15% of a full-time employee’s time.

What used to take 70 hours every year, now takes 2 hours, which frees up 15% of a full-time employee’s time. The client has been very happy with the results and is excited to convert other models to Modelshop as part of their digital transformation initiative. Based on CSG’s unique knowledge of the client’s activities, the project has given Mark and CSG the opportunity to provide a valuable service to the client, generating substantial savings and automating a number of key financial management functions.

Looking forward, Mark sees opportunities to create applications like this for other CSG clients. He has started to evaluate Modelshop’s APIs to provide a direct connection to data providers and is looking at Modelshop’s AI tools to more effectively automate manual processes while reducing client risk.

The solutions they deliver to their clients are much more collaborative, enabling greater immediacy and financial control for the clients.

Using Modelshop, CSG Advisors has the opportunity to provide highly valuable, solution-based offerings, precisely tailored to clients’ needs. The online applications they deliver to their clients are much more collaborative than the spreadsheets they are replacing, enabling greater immediacy and financial control for the clients and allowing Mark and his team to deliver a more attentive level of service.

Targeted Credit Offers using AI

Models that leverage machine learning and AI can predict, with greater accuracy, applicant behavior. It gives organizations an edge; they can offer products to applicants that they are more likely to purchase. This leads to higher conversion rates and more closed deals.

Watch Modelshop’s head of product create a targeted credit model live.

Hello World AI risk demo Watch the Demo