Introduction
In our ongoing exploration of AI's transformative role in the mortgage industry, we've been delving into the capabilities of various document automation solutions. In the first part of this series, we focused on a detailed comparison between Vaultedge and AWS API, shedding light on their respective strengths and weaknesses. Now, we're turning our attention to another key player in this evolving landscape: Google Lending API.
Google, a global technology giant renowned for its innovative use of AI, has entered the mortgage tech realm with the Google Lending API. At the same time, Vaultedge is making remarkable progress by utilizing AI to automate the entire document processing lifecycle. This raises an interesting question: how do these two impressive solutions compare?
In this second installment of our series, we'll conduct an in-depth comparison of Vaultedge and Google Lending API. We will scrutinize their features, assess their performance, and evaluate their potential impact on the future of loan document processing in the mortgage industry.
Our analysis will span from indexing to data extraction efficiency, from document processing speed to user experience. We aim to provide you with a comprehensive, fact-based perspective on these two powerful tools, without invoking any specific tool or mentioning its use.
Drawing from a range of sources, including head-to-head comparisons by industry experts like Murali Tirupati1, insights from Vaultedge's own website2, and details from Google's product page3, we will present a detailed, authoritative analysis. This will not only help industry insiders make informed decisions but also provide general readers with a clear understanding of the current AI landscape in the mortgage industry.
Software Overview :
Google Lending API - Document AI for Lending
The Google Lending API offers a transformative solution for borrowers and lenders, automating the mortgage document processing to enhance the home loan experience. By reducing processing time and streamlining data capture, this API enables faster loan processing while ensuring regulatory compliance.
Key Benefits:
- Automated data capture to expedite the home loan process from weeks to days.
- Increased operational efficiency in the loan process.
- Accelerated mortgage workflow processes, enabling seamless loan processing and automated document data capture with accuracy.
- Improved home loan experience by simplifying document process automation.
- Streamlined mortgage applications across all stages of the mortgage life cycle, resulting in faster time to close.
- Compliance support to enhance your regulatory posture, including data access controls, transparency, data residency, and customer-managed encryption keys.
Experience the power of the Google Lending API and revolutionize the way you handle mortgage document processing.
Document Types Recognized by Analyze Lending
The table below provides a comprehensive list of document types acknowledged by Google Document AI.
Vaultedge Document AI :
Vaultedge is a comprehensive end-to-end loan processing solution equipped with advanced document management features. It leverages the power of artificial intelligence and machine learning to extract, classify, and validate data from various types of mortgage documents. With its powerful OCR capabilities and custom-built models, Vaultedge accurately extracts data from structured and unstructured documents. This cutting-edge document AI software is purpose-built for the lending industry, utilizing large language models trained on millions of mortgage documents. Vaultedge not only accurately classifies documents but also extracts valuable information, making it an invaluable tool for streamlining lending processes.
Document Types Recognized by Vaultedge document AI
Vaultedge Document AI can accurately recognize and categorize over 500 mortgage document types and extract more than 2000 fields of information.
Side by Side Comparison
Performance Test:
We will perform a comparative analysis of the Google Lending API and Vaultage Document AI for the processing of lending documents. Our goal is to assess the capabilities of these solutions in parallel. We will utilize standard documents, including a note, deed of trust, and loan estimate, and evaluate the results obtained from each option.
Google Lending API - Document AI for Lending
Google APIs do not provide an option to upload these documents for classification purposes through the UI or console. However, we have processed these files and I would like to share with you the results identified by Google Lending Classifier.
Classification Results -
When we uploaded the documents to the Google Classifier in Google Lending API, Google indicated that the first five pages were classified as mortgage statements, while the next six pages were classified as a rider, and the last four pages as a HUD. Interestingly, all the pages were classified as mortgage statements when it came to the note. The confidence levels of these classifications were also visible. Hence, it appears that the classification by the Google Classifier is inaccurate.
Extraction Results -
Now, let's examine the data extracted by Google. As you can see, Google has extracted the data mentioned here. This is a note. Upon closer inspection, you will observe that the software has picked out relevant information such as state, city, and property address. However, it is important to note that the software selects this information based on its own judgment without accurately identifying the specific data. For instance, if we focus on the paragraph "Borrower's promise to pay," our main interest lies in extracting details about the principal and the lender. Ideally, we would expect to extract the principal's name and the lender's name from that paragraph. Unfortunately, the current output provides the entire paragraph without any specific data extraction.
Upon examining the output from the Google Lending API for a deed of trust, we once again observe a recurring pattern. It is evident that the software extracted both the MIN and MERS phone numbers without making any distinction. However, it is notable that the software does not identify the extracted number as a MIN number. Instead, it faithfully extracts the relevant paragraph from the document.
Even upon examining the output generated by the Google Lending API for a deed of trust, we observe a recurring pattern. The software accurately extracts the MIN and MERS phone numbers, however, it fails to provide any distinction. It does not specify that the extracted number is a MIN number. Instead, the software verbatim extracts the paragraph it deems relevant from the document.
If I were to explain the functionality of both the Google Lending API and AWS Analyze Lending APIs, they perform adequately when you upload standard format documents such as a W2 or a 1040. However, if you need to upload an entire loan file containing multiple documents, including unstructured and semi-structured ones like a note, deed of trust, or insurance, the results are not satisfactory for practical use.
In Vaultedge, we have the convenience of uploading all documents at once, eliminating the need to upload them one by one as in AWS Lending API.
Now, let's examine the indexing output. During this process, the software accurately identified the first three pages as a loan estimate, followed by the next 16 pages categorized as a mortgage deed of trust. Finally, the last three pages were identified as a note.
Moving on to data extraction, Vaultedge impressively extracted a significant amount of information from the deed of trust, including the property address, MERS ID, borrower, co-borrower, and more.
The software also provides a confidence rating, allowing for efficient review of low confidence items. By leveraging Vaultedge, time is saved by focusing only on the two pieces of data flagged as low confidence.
Similarly, for notes, Vaultedge extracts a wealth of information, providing the option to review low confidence items if desired. This stands in contrast to AWS Lending API, which failed to extract the interest rate.
Vaultedge's accuracy and comprehensive data extraction capabilities make it a valuable tool for mortgage document processing.
Conclusion
Based on my comparison, when explaining the functionality of both the Google Lending API and AWS Analyze Lending APIs, they demonstrate satisfactory performance when uploading standard format documents such as a W2 or a 1040. However, if there is a need to upload an entire loan file comprising multiple documents, including unstructured and semi-structured ones like a note, deed of trust, or insurance, the practical usability of the results is not satisfactory.
On a final note, if you are in the mortgage industry and exploring high accuracy document automation solutions for semi-structured & unstructured data, then let's chat to see what could be better alternatives to your current approach.
Schedule a call with us today and let's find the best solution for your mortgage document needs!