Capture Uses AI-Assisted Data Collection for 69% of Required Fields

CAPTURE Data Services, LLCFounded by Jon Forrester with his 25+ years of appraisal and valuation experience, CAPTURE Data Services is designed to reinvent the way that property data is collected.

Capture Case study statistics


Jon’s background in the appraisal space provided him a first-hand look at the challenges of collecting data about a home. While an appraiser or inspector is on-site, there are a multitude of details they must manually record in what becomes an expensive and time-consuming process. In the event certain details are missed, even more money and time are wasted revisiting a property.  


Jon’s long-term vision for CAPTURE was for a human to walk through a property with a mobile device and document the space by taking photos. AI would scan each photo and based on its analysis, property details would be automatically recorded with the user prompted for additional information where needed. Having heard about Restb.ai’s computer vision solutions, CAPTURE reached out to see if their AI solutions could help test this theory. 

CAPTURE app menu
CAPTURE Kitchen Overview Screen


CAPTURE and Restb.ai decided to roll things out in a two-phase approach. First as a quality control check on completed property data collection (PDC) assignments, and secondly as a key part of an integrated workflow.


Initially, Restb.ai worked as a validation process to ensure that all recorded features were correctly populated. CAPTURE estimated that 69% of the details collected in their app could be identified from photos and these items could be automatically reviewed to ensure the user input details matched what was visible in the photos. This process required minimal development effort as it was built upon the CAPTURE PDC app’s existing workflow and designed to not require any UX changes.


As CAPTURE grew comfortable with the AI’s performance, they began to see the value that computer vision provided with each design change they had moving forward. Over time, CAPTURE leveraged Restb.ai’s solutions to dynamically alter the app’s workflow with each photo taken. Depending on the contents detected in an image, details are automatically populated within the app and the user is prompted with different questions as needed. 

“Restb.ai is essential for my vision of data collection. Leveraging AI I can improve the quality of data captured, increase the amount of data collected and decrease the amount of time spent on site.” 

Jon Forrester – President at CAPTURE Data Services 


After working together for almost a year, CAPTURE has been able to benefit from Restb.ai’s technology in three key areas: Efficiency, Quality, and Completeness.



CAPTURE prides itself on the industry’s fastest and most reliable property data collection process. Restb.ai’s auto-population of fields saves the CAPTURE community valuable minutes off their data collection time, while creating a more user-friendly experience. Using a standardized AI model to scan images is particularly important for helping confirm and standardize a home’s condition, which can be subjective and susceptible to sub-conscious biases. Lastly, it ensures that all features are accurately captured in the photos to help prevent having to revisit a property due to missed information.



Restb.ai’s solutions were also able to increase the overall quality of reports taken. Automated quality checks ensure that every report can be instantly reviewed. Furthermore, CAPTURE implemented Restb.ai’s feedback system that allows the models to learn where there are disagreements between the AI and the user. Over the past 6 months, Restb.ai has been able to continually iterate to improve CAPTURE’s accuracy and reduce the number of AI errors by 53%. Even minor “more cosmetic” items like having a user’s finger partially covering an image, can now be eradicated thanks to AI’s help.



Lastly, Restb.ai has helped CAPTURE enrich the data that is collected. Many PDC reports are focused on adapting to different standards, such as those required for hybrid or bifurcated appraisals. It is difficult to justify collecting details beyond what is required because it increases the amount of time the user spends at the property. With Restb.ai, CAPTURE was able to collect over 100 additional data points such as the color of appliances, the kitchen’s layout, the styling of each room and more – all in the same amount of time!


For more information on our image tagging solutions or CAPTURE Data Services, please visit...