LeaseQuery AI
​
Leveraging AI to Automate Lease Agreement
Processing for Lease Entry
My Role: UX Designer | Duration: 1 Year | Project Status: Completed
In the realm of finance, lease agreements can be complex documents that often require meticulous review and data extraction for accurate lease accounting.
The objective of this project was to develop an AI-powered tool that automates the extraction of key information from lease agreements, streamlining the process of filling out lease accounting forms.
Lease accountants spend a significant amount of time manually reviewing lease agreements and inputting data into accounting systems. This manual process is not only time-consuming but also prone to human error, which can lead to financial discrepancies and compliance issues. There was a clear need for a solution that could reduce time spent on these tasks while improving accuracy.
What do we need to know?
We conducted interviews with lease accountants and stakeholders to understand their workflows and pain points. We also gathered data from other sources, such as, Productboard, Fullstory, Heap, NPS, and more.​
​​​​
​
Our Biggest Takeaways​
​
-
Context is key.
-
AI shouldn't get in the way of the normal process; it should enhance it.
-
Pre-filled fields don't immediately equal faster lease entry.
-
Accuracy can only take you so far.​
-
Lease accounting knowledge is still important.
-
-
How can we provide value even when the AI is inaccurate?
​
​
Meet the User
To encapsulate our research and guide our design process, we developed three user personas, which visually represent our target audience.
​
The Problem
Lease accountants spend a significant amount of time manually reviewing lease agreements and inputting data into accounting systems. This manual process is not only time-consuming but also prone to human error, which can lead to financial discrepancies and compliance issues. There was a clear need for a solution that could reduce time spent on these tasks while improving accuracy.​
The Goal
​We need a design that accommodates potential inaccuracies as the machine continues to learn, addresses the time needed for generating predictions, and considers the scope of its predictive capabilities.
Introducing LQ AI: Streamlining Lease Entry for Accountants
With LQ AI, you can now upload your lease documents, and our advanced AI will automatically read the details and fill in your lease entry forms for you.
​
LQ AI intelligently extracts key lease terms, dates, payment details, and other critical information from your uploaded lease agreements, reducing errors and saving you valuable time. Whether you’re handling a single lease or processing multiple agreements, LQ AI is here to help you work smarter, not harder.
​
Key Benefits:
-
Faster Processing: Quickly convert lease documents into accurate lease entries.
-
Error Reduction: Automated data extraction minimizes the risk of human error.
-
Seamless Integration: Easily upload your documents and let LQ AI do the heavy lifting.​
​
With LQ AI, accountants can focus more on strategic tasks and less on manual data entry, all while ensuring accuracy and compliance.
​​
1
Uploading PDF
Users can easily upload their documents using the new document uploader, located on the right side of the lease entry section. Once the document is uploaded, the AI can initiate its scanning process, which usually involves Optical Character Recognition (OCR)
2
Viewing Predictions
Once the PDF has been scanned, fields with available predictions will be highlighted with an orange sparkle icon (our AI symbol).
Users can simply click on these highlighted areas to view the AI-generated predictions directly within the document.
A collapsible sidebar featuring predictions, allowing you to effortlessly view all predictions at a glance.
2
Continued: Various Types of Predictions
Some fields require more than just a simple prediction. To address this, we've developed AI Summary and AI Group predictions. These features allow the AI to automatically fill in multiple fields with a single prediction, streamlining the process for more complex scenarios.
Users can resize the PDF viewer by hovering over the edge until the resize cursor appears, then dragging left or right.
3
Managing PDF Viewer
User Testing
Throughout the design process, we conducted extensive user testing to ensure our application was both intuitive and easy to navigate. This included a mix of moderated Zoom sessions and unmoderated Maze surveys with real users, allowing us to gather valuable feedback. Our primary goal was to make sure the average user could seamlessly interact with our new AI features without disruption.
​
We tested key elements, including the placement of our AI icon (the signature orange sparkle), how to introduce the AI product in a way that felt natural and non-intrusive, and how users could engage with predictions. This involved testing how they could view, select, or dismiss predictions and even manually input their own data. Every decision was driven by user insights to create a smoother, more efficient experience.
​
The Outcome
The LQ QI was released to beta users, who were monitored through Heap and FullStory as they worked through their workflows. All participants successfully completed their tasks without any issues. Additionally, we surveyed them to assess their understanding of the available features and how to use them. The feedback was overwhelmingly positive, with users finding the features intuitive and easy to navigate.​
Next Steps
The next steps involve continuing to refine and enhance the machine learning capabilities of our AI, ensuring that our design evolves in tandem with its growing abilities. We'll remain committed to ongoing user engagement, gathering valuable feedback, and analyzing behavioral data from Heap and FullStory. This will help us identify opportunities to further optimize the user experience and ensure our design is intuitive, efficient, and responsive to user needs. Our goal is to continually adapt the design to maximize usability while leveraging the AI’s full potential.​
​​
​