Rogo delivers secure AI with Amazon Bedrock, driving innovation in finance

Wall Street stands at a pivotal moment in its technological evolution. While generative artificial intelligence (AI) promises to revolutionize how financial professionals work, the industry’s stringent compliance requirements and complex legacy systems have created significant adoption challenges for firms. Many banks still rely on siloed core systems that are decades old, making it difficult and expensive to integrate new tech.

Yet the potential in marrying these industries is undeniable —Gartner reports that 70 percent of financial services leaders believe that adopting generative AI would be beneficial for their organization. By 2026, it’s estimated that finance organizations with over three years of AI skills investment will double their productivity over organizations with no AI skills. With this, enter Rogo: an innovative startup that’s bridging this critical gap.

Rogo’s generative AI platform is purpose-built to help finance professionals conduct research, make data-driven decisions, uncover deal-winning insights, and produce essential materials in minutes—not days. Together with Amazon Web Services (AWS), the company is helping shape the future of finance for world’s largest banks and private equity investors with a secure, compliance-driven AI platform.

From Investment Banking to Building an AI Start-Up

Rogo was co-founded by Gabriel Stengel (CEO), John Willett (COO), and Tumas Rackaitis (CTO). Stengel and Willett developed the earliest version of Rogo while studying together at Princeton. After working at Lazard and J.P. Morgan respectively, they each recognized the immense opportunity in what it would mean to apply AI within the day-to-day work of bankers. The pair later teamed up with Tumas Rackaitis, a software engineer (and trader at a New York City-based hedge fund,) to help build the business.

Today, Rogo’s platform comprises three key offerings:

1. A highly accurate and auditable AI analyst with search and chat capabilities, powered by fine-tuned large language models (LLM). 
2. A framework of AI agents for applications like PowerPoint and Excel, delivering out-of-the-box automation for time-consuming tasks like presentation slide creation. 
3. A buy-and-build AI strategy, in which Rogo experts provide white glove support and infrastructure to businesses developing and implementing their own AI strategy.

“Our platform provides immediate benefits through an out-of-the-box, turnkey solution that enables finance professionals to process more deals, expand coverage, and get smart faster—all in a secure environment built for the financial industry’s complex demands,” says Rackaitis. Through Rogo, firms can securely integrate their internal data, as well as a library of over 65 million high-quality sources, including market research reports, Securities and Exchange Commission (SEC) and international filings, transcripts, live news, and private company information. In a single platform, firms can analyze datasets to generate instant, accurate answers, supported by auditable citations.

For instance, a senior figure working at a leading financial firm can use Rogo’s AI analyst to find accurate answers to queries in real-time. “If you’re speaking with a client and they ask about recent deal activity, you can ask Rogo and instantly receive information you can trust, as well as gain novel insights,” says Rackaitis. “We’ve done extensive user research and immense amounts of empirical science to make sure that our customers can trust the LLM output, that we don’t hallucinate, and that it’s easy to audit the source of information.”

Customers can also create highly specialized workflows that save time by accelerating common tasks - from creating presentation slides and preparing for meetings, to summarizing documents and generating well-researched reports in hours (that typically take analysts weeks to produce.) Crucially, Rogo AI is designed to converse with users in a clear, understandable manner without getting bogged down in technical jargon. “We’ve had a lot of users point to that educational, informative, and journalistic style as really, really helpful,” says Rackaitis.

Collaboration that compounds growth

For Fintech startups to succeed, they must iterate and innovate at pace to outcompete current incumbents, while also meeting stringent regulatory, cybersecurity, and compliance requirements. This tension is felt even more strikingly in the generative AI space, where new models and technical advancements are occurring at an extraordinary rate. “AWS contains all the tools an AI startup needs to design, scale, and secure their product. The reporting, auditability, and governance controls meet our strict security requirements,” says Rackaitis.

Andrew Canning, Lead Solutions Architect at Rogo, adds: “Financial firms are hungry to be on AWS themselves if they’re not on it already. It’s essentially an industry standard. AWS has the widest range of security services that everyone trusts. Security teams working at financial firms are already familiar with a lot of AWS services. If I say to them that we’re going to use Amazon EKS, or AWS Certificate Manager, or AWS Route 53, they’re already on board.”

Amazon Bedrock—velocity meets security

Generative AI is evolving rapidly, and AI startups that are able to access high performing foundation models (FMs) can gain a competitive edge, faster. Rogo is using Amazon Bedrock to do just that. A fully managed service, Amazon Bedrock delivers access to the latest FMs from leading providers—all through a single API. This enables businesses to easily evaluate, experiment with, and privately customize FMs based on their intended use case.

Amazon Bedrock is also serverless, so technical teams don’t have to manage any infrastructure, and can securely integrate and deploy generative AI capabilities into their applications using the AWS services they are already familiar with. This made it a natural choice for Rogo. “We were already on AWS, and Amazon Bedrock allowed us to stay within that deployment and take advantage of a private instance of an Anthropic model,” says Rackaitis. “Those models are exceptionally good and Amazon Bedrock allows us to work with them on private and secure compute.”

In addition, Amazon Bedrock is equipped with deep security features spanning everything from full data encryption in transit and at rest, to robust logging, monitoring, and access management capabilities. “Amazon Bedrock allows us to protect sensitive information,” says Rackaitis. “We got access to the tool early. We’re one of the first alpha users, and it was really cool to work so closely with the AWS team from day zero. It shows how AWS prioritizes startups.”

“We always ask ourselves: How can we guarantee where data goes? How can we guarantee isolation on a network level and a hardware level? A bank for example is essentially the assets on its books, and its books are files. There are many different types of files and modalities that are unique to finance. Dealing with financial Excel sheets and slides for example is exceptionally difficult, and we spent a lot of R&D effort and resources building our own file ingestion system,” says Rackaitis. “That was the biggest rationale behind choosing Amazon Bedrock—we get velocity and security all in one.”

When a foundation model is tuned on Amazon Bedrock, it is based on a private copy of that model—data is never shared with model providers and is not used to improve the base models. It’s also easy to establish private connectivity between Amazon Bedrock and an Amazon Private Cloud (VPC) without exposing that VPC to internet traffic. That functionality is invaluable for a company that offers a range of secure deployment options.

Rogo customers can choose between complete data isolation in which customers are given their own Kubernetes clusters and data stores, or account-level isolation wherein additional networking and other security layers are provided on top. “It all comes down to compliance conversations,” says Canning. “Customers can dictate how they want to deploy based on their needs which are unique to each firm.”

Banking on future growth

Rogo continues to work closely with AWS to take advantage of new technology and capture new business opportunities. “Whenever we’ve had questions about a specific AWS service, we’ve been able to have really exceptional conversations with ex-startup founders and other experts at AWS,” says Rackaitis.

Collaboration with AWS is also proving valuable in quickly earning the trust of prospective and existing customers. “It’s really awesome walking into compliance meetings all day, speaking to legacy firms with sprawling security teams, and having AWS at our back,” says Canning. “We can confidently say to these firms that we have a platform that is on par with theirs in terms of security posture—better in some circumstances.”

Going forward, Rogo is committed to exploring the boundaries of FinTech’s AI frontier with AWS. “We’re exploring a lot of different AI models. We’re investing heavily in reinforcement learning which I think is going to be a major differentiator for us. That’s going to involve working more closely with AWS teams on things like GPU choice and other computing paradigms,” says Rackaitis. “We’re also looking to standardize a lot of the white glove customization we offer for security and make that available for all customers.”

Canning adds: “From a security perspective, I’m really excited about AWS Nitro Enclaves, which will allow us to essentially lock ourselves out of our own environment and give customers the keys to securely access their data with complete ownership. It’s a hot topic in the financial space, and I believe it’s going to really help Rogo’s security capabilities reach the stratosphere.”