Revolutionizing Real Estate Transactions Using AI on AWS with Rexera

Rexera provides AI-powered solutions that address key issues in the real estate industry—for example, acquiring and processing high volumes of legal documents, extracting critical information from those documents, and flagging potential obstacles. These solutions also help minimize manual errors that can have significant legal and financial consequences, in addition to identifying counterparties such as municipalities, homeowners associations, lenders, and utility companies. “Our solutions use AI to reduce human involvement so that homeowners can buy and sell quickly with much less friction and at lower costs,” says Malhotra. “Furthermore, real estate professionals can spend more time growing their business and less time chasing documents.”

The company began using AWS early on. “We knew from the beginning that we had to find the right cloud provider because we wanted to be able to scale aggressively as we grew,” says Malhotra. “The early setup on AWS was straightforward, and as we scaled, we expanded the number of services that we use.”

One service that Rexera adopted as it grew was Amazon Bedrock, a fully managed service that offers a choice of high-performing foundation models from leading AI companies. The company migrated to Amazon Bedrock from OpenAI. Amazon Bedrock also connected seamlessly to the rest of Rexera’s AWS environment, so the company was able to smoothly incorporate generative AI capabilities into its existing workflows. “On Amazon Bedrock, we could have more control and stability for our operations,” says Malhotra. After a 2-week proof of concept and a 2-week migration, Rexera moved fully into production on Amazon Bedrock in the second quarter of 2024.

Rexera uses large language models in Amazon Bedrock to process 5 million pages of documents each month. Thus, the company reduces manual document-processing workloads for customers by 99 percent and minimizes document-related risks and inaccuracies. On average, Rexera’s solution saves customers 4 hours on manual document review in each transaction and lowers operational costs by 25 percent. “Using Amazon Bedrock, we can process complex data types, deliver fast results, and improve overall output,” says Malhotra.

Rexera uses Amazon Bedrock to select the perfect AI models for each task, including  powerful options like Meta's Llama  and Anthropic’s Claude. This variety helps the company enhance its solution and reduce biases and hallucinations. “With a multiple-model approach, we have an accuracy rate of more than 95 percent and need less human intervention in straightforward cases,” says Malhotra. “This helps our team significantly increase processing capacity and service more transactions.”

Processing large, unstandardized documents is a significantly time-consuming task in real estate. To extract key details from hundreds of pages in seconds, Rexera built a tool that uses retrieval-augmented generation—a generative AI architecture that uses specific context data, in this case, real estate documents such as homeowners associations’ agreements and property records. This tool reduces manual review time by 80 percent and significantly lowers the risk of missing critical information.

Rexera’s solutions free up real estate professionals to focus on their clients rather than spending time on inefficient manual processes. Rexera’s AI agents can meticulously verify details in customer communications, run quality control across transactions to flag potential errors or issues, and even call entities such as homeowners associations to confirm property information and request documents. With these services, Rexera’s customers decrease risks and costs. “Because Rexera takes over manual, error-prone work that no one likes doing, our customers can finally focus on growing their business,” says Xenia Cotlearova, marketing director at Rexera.

Rexera uses the flexibility of Amazon Bedrock to develop, test, and deploy new features and improvements quickly. “Using Amazon Bedrock, we are years ahead of where we initially planned to be on our product road map,” says Malhotra. “Our plan is to continue to build additional agentic workflows for our customers to minimize friction in real estate and reduce the costs, time, and complexity of transactions.”

Rexera is also becoming an AWS Partner to continue expanding its solution and customer base. For other startups, Malhotra recommends being proactive in engaging with the AWS team. “It offers dedicated resources and guidance that are tailored to specific project needs, which was invaluable for us to migrate seamlessly and scale our workloads effectively,” says Malhotra.