Introduction
Thomson Reuters has long been a pioneer in delivering information and analytics to professionals across finance, legal, tax, and media. In recent years the company has turned its focus toward democratizing artificial intelligence, making it accessible to users who may not have a background in data science or software engineering. The result is Open Arena, a no‑code AI platform that sits on top of Amazon Bedrock and a suite of AWS services. The platform is designed to let business users build, deploy, and iterate on AI models without writing a single line of code, while still providing the robustness, scalability, and security that a global enterprise demands.
Open Arena is more than a set of tools; it is an ecosystem that integrates the power of Bedrock’s foundation models with the data‑handling capabilities of Amazon OpenSearch Service, Amazon S3, Amazon DynamoDB, and AWS Lambda. By combining these services, Thomson Reuters has created a workflow that supports everything from data ingestion and preprocessing to model inference and continuous monitoring. The platform’s architecture is intentionally modular, allowing organizations to plug in new data sources or replace components as their needs evolve.
The goal of this post is to walk through the key architectural decisions, illustrate how the platform solves real‑world business problems, and highlight the types of professionals who benefit from Open Arena. Whether you are a compliance officer looking to automate document review, a financial analyst seeking to generate market insights, or a legal researcher wanting to surface precedent, the platform’s no‑code interface and underlying AWS infrastructure make it possible to bring AI into everyday workflows.
Main Content
Architectural Foundations
At the heart of Open Arena is a microservices‑based architecture that leverages AWS Lambda for compute, Amazon DynamoDB for low‑latency metadata storage, and Amazon S3 for durable object storage. Lambda functions orchestrate data pipelines that feed into Bedrock’s foundation models. Because Lambda scales automatically, the platform can handle spikes in demand—such as a sudden influx of regulatory filings—without manual intervention.
The data layer is built on Amazon OpenSearch Service, which provides full‑text search, analytics, and real‑time indexing. By indexing processed documents and model outputs in OpenSearch, users can perform ad‑hoc queries, generate dashboards, and surface insights directly from the platform. This eliminates the need for separate BI tools and keeps the entire workflow within a single, secure environment.
Security and governance are baked into every layer. All data at rest is encrypted using AWS Key Management Service (KMS), and all traffic between services is encrypted in transit using TLS. Role‑based access control is enforced through AWS Identity and Access Management (IAM), ensuring that only authorized users can create or modify models, view data, or trigger inference jobs.
No‑Code AI with Amazon Bedrock
Amazon Bedrock provides access to a library of pre‑trained foundation models from leading vendors, including OpenAI, Anthropic, and Stability AI. Open Arena exposes these models through a visual interface that lets users define prompts, set parameters, and specify output formats without writing code. Users can drag and drop components to build pipelines: for instance, a document ingestion node can feed into a summarization model, which then routes the output to an OpenSearch index.
The platform also supports fine‑tuning on proprietary data. Users can upload a small set of domain‑specific documents to S3, and Lambda functions will orchestrate the fine‑tuning process on Bedrock. Because the fine‑tuning job runs entirely within AWS, there is no need for on‑prem infrastructure or data exfiltration, preserving compliance with data residency regulations.
Integrating AWS Services for Scalability
Scalability is achieved through a combination of serverless compute, managed storage, and event‑driven architecture. When a user uploads a new document, an S3 event triggers a Lambda function that extracts metadata, stores it in DynamoDB, and indexes the content in OpenSearch. The same function can optionally invoke a Bedrock inference job to generate a summary or extract entities.
For high‑volume scenarios, such as processing thousands of legal briefs per day, the platform can spin up additional Lambda instances automatically. Because Lambda functions are stateless, they can be distributed across multiple availability zones, providing fault tolerance and reducing latency.
Monitoring and observability are handled by Amazon CloudWatch. Metrics such as request latency, error rates, and resource utilization are exposed to users through a custom dashboard. This transparency allows business analysts to identify bottlenecks and optimize their workflows without needing to dive into the underlying code.
Business Use Cases
Open Arena has been deployed across several verticals within Thomson Reuters. In the legal domain, compliance teams use the platform to automatically review contracts, flag non‑compliant clauses, and generate summary reports. The no‑code interface allows paralegals to set up a pipeline that ingests new contracts, runs them through a fine‑tuned Bedrock model, and stores the results in OpenSearch for quick retrieval.
In finance, analysts leverage Open Arena to scan earnings call transcripts, extract sentiment scores, and correlate them with market movements. By integrating with Amazon OpenSearch, analysts can run real‑time queries that surface the most influential statements, enabling faster decision‑making.
Tax professionals use the platform to parse complex regulatory documents, identify key changes, and generate compliance checklists. The fine‑tuning capability ensures that the model understands the nuances of tax language, reducing the risk of oversight.
Across all use cases, the common theme is that users can prototype and deploy AI solutions in days rather than months, freeing up data scientists to focus on higher‑value tasks.
User Profiles and Adoption
Open Arena is designed for a broad spectrum of users. Technical teams—data scientists, ML engineers, and DevOps—can use the platform to experiment with new models, monitor performance, and integrate the results into existing pipelines. Non‑technical users—business analysts, compliance officers, and subject‑matter experts—benefit from the visual workflow builder, which abstracts away the complexity of model configuration.
Adoption has been accelerated by Thomson Reuters’ internal training programs, which provide hands‑on workshops and documentation. The platform’s API endpoints also allow external partners to embed AI capabilities into their own applications, expanding the reach beyond the company’s internal teams.
Conclusion
Open Arena represents a significant step forward in making artificial intelligence accessible to professionals who do not have a technical background. By leveraging Amazon Bedrock’s powerful foundation models and the scalability of AWS services, Thomson Reuters has built a platform that is both robust and user‑friendly. The architecture’s modularity ensures that it can evolve alongside emerging AI capabilities, while its security posture meets the stringent requirements of regulated industries.
The platform’s impact is already evident in the speed with which business units can prototype, deploy, and iterate on AI solutions. From automating contract review to generating market insights, Open Arena is turning AI from a niche capability into a mainstream business enabler. As AI continues to permeate every facet of professional work, platforms like Open Arena will play a pivotal role in ensuring that the benefits of advanced analytics are available to all.
Call to Action
If you are part of an organization looking to unlock the power of AI without building a data science team from scratch, consider exploring Open Arena. Reach out to Thomson Reuters’ AI enablement team to schedule a live demo, or sign up for a free trial to experience the no‑code workflow firsthand. By embracing a platform that combines the flexibility of Bedrock with the reliability of AWS, you can accelerate innovation, reduce time‑to‑value, and empower your professionals to make data‑driven decisions with confidence.
