Building a Remote Deposition Solution Using AWS Fargate and .NET with Remote Legal

Remote Legal offers a solution that attorneys can use to host remote depositions and legal proceedings online. Remote Legal launched in 2021 after its parent company, Precision Reporters, saw a gap in remote conferencing technology for attorneys at the beginning of the COVID-19 pandemic. Remote Legal built a solution that put attorneys first—offering features such as instant exhibit uploading, live exhibit marking and annotating, picture-in-picture video, and instant voice-to-text transcription.

Remote Legal built its solution in 10 months using ASP.NET Core SignalR for near-real-time web functionality in addition to managed services on AWS. ASP.NET Core SignalR is an open-source, cross-platform library which can run on Windows, macOS, and Linux. Remote Legal chose to run ASP.NET Core on Linux for optimal cost and performance. To host ASP.NET Core SignalR applications, Remote Legal used AWS Fargate—a serverless, pay-as-you-go compute engine—and Application Load Balancer, which performs advanced load balancing by routing traffic to targets based on the content of the request.

The solution also made use of Amazon Relational Database Service (Amazon RDS) for MySQL, a collection of services that makes it simple to set up, operate, and scale databases in the cloud. The company released its product in May 2021. Remote Legal wanted to use serverless architecture and managed services to maintain a small footprint and to efficiently allocate its limited resources. “The beauty of using a cloud provider such as AWS is the scalability,” says Alex Rodriguez, senior director of technology at Remote Legal. “We’ve seen it already. As we grow and as our solution is increasingly adopted, it’s simple to expand and add additional services.”

Remote Legal’s solution has three core elements: video, document management, and transcription. In April 2023, Remote Legal began experiencing errors in its cloud transcription service. Its service provider had updated its large language model, which resulted in persistent phantom text errors—words and sentences that had never been said appeared in the transcripts after batch processing. This issue was not resolved, so the Remote Legal team decided to migrate to a new transcription service. The team investigated multiple solutions but ultimately chose AWS for its range of services, strong security measures, cost-effectiveness, and customer service.

The Remote Legal team chose Amazon Transcribe, a fully managed speech recognition service that automatically converts speech to text, to replace its previous transcription service. In 1 week, the team completed the migration using AWS SDK for .NET—which makes it simple to call AWS services using idiomatic .NET APIs—and virtually removed transcription errors. Remote Legal uses many other AWS services to power its solution, and the team appreciated the seamless integration of Amazon Transcribe with these services. “AWS services are like building blocks,” says Rodriguez. “It’s simple to put them together and make them work.”

Remote Legal’s entire architecture on AWS is infrastructure as code, which makes the solution extremely consistent and stable. Not only can the team quickly create new environments for testing or creating proofs of concept within virtually any region, but it can also restore identical configurations in new regions in the event of a disaster. With such stability, the solution reached a 98 percent success rate, meaning that the depositions that start on the solution finish on it. “If a deposition starts on our solution but then moves to another, that’s a bust,” says Rodriguez. “We’ve stayed under five busts per month throughout the year because of the stability of our solution and quality of our code, alongside the availability of AWS infrastructure.”

On AWS, Remote Legal also generated cost savings. Each month, the Remote Legal team meets with Rapidscale, an AWS Premier Tier Services Partner, to review costs and identify opportunities to optimize them. Rapidscale recommended automating Remote Legal’s testing and demo environments. These environments, which were previously running all the time, began to be operated only when the development and sales teams used them. This reduced Remote Legal’s costs by 85 percent compared with the cost of the production environment.

Other cost optimizations have reduced operational overhead for Remote Legal. “We have no DevOps team employed at Remote Legal,” says Rodriguez. “That’s all managed by AWS and our partner. We’re not responsible for server maintenance or applying patches, which is a huge responsibility. It would be a headache to manage everything alone.” By using managed services and outsourcing some work to its partner, Remote Legal estimates that it saves half the yearly salary of the cloud engineer it would otherwise employ.

Using fully managed services on AWS, the Remote Legal team can focus on optimizing and improving its services. Now that Remote Legal has added artificial intelligence (AI) using Amazon Transcribe, next up is adding generative AI capabilities to its solution. The team hopes to achieve even higher customer satisfaction by using AI to identify and solve issues such as echoing or lagging during depositions. Moreover, Remote Legal plans to use AI to create new products related to transcription.

“We strive to remove anything that adds friction to people’s experience of our product,” says Rodriguez. “Benefiting from the power and reliability of AWS, we offer a frictionless experience to our customers.”