LIVE
OpenAI Partners with Chilean Ministry to Advance Public Sector AI Ambience Healthcare Expands Clinical AI Suite for Hospital Operations AI Models Vulnerable to Manipulation in Simulated Warfare Scenarios, Research Shows AWS Highlights Enterprise Governance and AI Workspace Quick Ahead of re:Invent 2026 LoanPro Integrates Amazon Bedrock to Automate Loan Servicing Workflows Google Brings Wallet App to iOS Users Across the United States Microsoft Reveals Cause of Global Outlook Outage AWS Outlines Phased Roadmap for Responsible Agentic AI Governance Google updates Gemini settings to separate chat history from AI model training Uber Eats and Zipline launch autonomous drone food delivery in Houston OpenAI Partners with Chilean Ministry to Advance Public Sector AI Ambience Healthcare Expands Clinical AI Suite for Hospital Operations AI Models Vulnerable to Manipulation in Simulated Warfare Scenarios, Research Shows AWS Highlights Enterprise Governance and AI Workspace Quick Ahead of re:Invent 2026 LoanPro Integrates Amazon Bedrock to Automate Loan Servicing Workflows Google Brings Wallet App to iOS Users Across the United States Microsoft Reveals Cause of Global Outlook Outage AWS Outlines Phased Roadmap for Responsible Agentic AI Governance Google updates Gemini settings to separate chat history from AI model training Uber Eats and Zipline launch autonomous drone food delivery in Houston

ArisGlobal Cuts Generative AI Costs on Amazon Bedrock Using Prompt Caching

🔌 Chips & Hardware 2 min read 👁 5

Life sciences tech provider ArisGlobal scaled its generative AI engine NavaX on Amazon Bedrock by optimizing prompt caching strategies to manage rising infrastructure expenses.

ArisGlobal Cuts Generative AI Costs on Amazon Bedrock Using Prompt Caching🔌Chips & Hardware

Photo: AWS Machine Learning Blog

Key points

  • ArisGlobal scaled its life sciences AI engine NavaX using Amazon Bedrock.
  • Compliance additions caused shared prompt prefixes to expand by 150 percent over six months.
  • Reversing cache priorities from source documents to shared prefixes mitigated a 200 percent surge in per-case inference costs.

Life sciences technology provider ArisGlobal has optimized the inference architecture of its NavaX artificial intelligence engine on Amazon Bedrock, successfully curbing escalating operational expenses. NavaX, which automates adverse-event case processing for pharmaceutical companies, runs on Amazon Bedrock utilizing Anthropic’s Claude models. The system extracts roughly 130 data fields from complex, multipage medical documents to streamline drug safety workflows for enterprise clients.

As ArisGlobal onboarded new corporate clients, compliance mandates caused the shared prompt prefix—which contains system rules, extraction schemas, and few-shot examples—to expand by 150 percent over a six-month period, stretching from 72,000 to 180,000 tokens. Because the platform executes multiple sequential requests per case, this compounding prompt growth drove per-case inference costs up by more than 200 percent.

Initially, the system prioritized caching source documents, which left the expanding shared prompt prefix outside the cache and subject to repeated reprocessing. By reversing the request structure to prioritize the shared prefix for explicit cache checkpoints, ArisGlobal reduced effective input costs and successfully supported higher tenant volumes without incurring proportional financial growth.

What to watch next: Monitor how other enterprise SaaS providers implement prefix caching and explicit checkpoints on cloud-managed foundational models to control scaling expenses.

Source AWS Machine Learning Blog — aws.amazon.com/blogs/machine-learning/how-arisglobal-scaled-generative-… ↗

Contact the newsroom

Got a tip, a proposal or a partnership question? Write to us and we will reply shortly.