Bluesight, a healthcare software company, has launched Prism Assistant, an AI agent built on Amazon Bedrock that automates drug compliance audits across six hospital management products. The tool lets hospitals avoid manually cross-referencing drug purchases against FDA shortages, inventory data, and eligibility records—a task that consumes over 4,000 hours annually per hospital. The first version launched in May 2026 and is already in use by 20 health systems.
Summaries like this, in your inbox every morning.
Sign up free →What happened
Healthcare software company Bluesight, using Amazon Bedrock AgentCore, built Prism Assistant—an AI agent that reasons across multiple products (ControlCheck, CostCheck, ShortageCheck, 340BCheck, and others) to automate hospital compliance tasks. Prism Assistant for ControlCheck launched in May 2026 and is already in use by 20 health systems; a more complex multi-product version is on track for later in 2026.
Why it matters
Hospitals managing drug pricing compliance currently spend over 4,000 hours annually per facility manually cross-referencing purchases against FDA shortage lists, inventory data, and backorder signals—work that scales poorly across networks of hundreds of hospitals. Prism automates this by letting an AI agent query data from multiple systems at once and surface actionable insights without manual report compilation, addressing a long-standing customer request that cut across product boundaries.
What to watch
The solution was built in a three-day AWS sprint in September 2025 and moved to production in under nine months—a timeline that would typically take 12–18 months. The architecture separates AI reasoning from the data layer, wrapping existing APIs in AWS Lambda to reduce query latency from 5 minutes to 10 seconds. The second multi-product version is expected later in 2026.
Bluesight's move toward agentic AI reflects a broader shift in healthcare compliance: the scale of manual work has become untenable. Hospitals classified as DSH (Disproportionate Share), PED (Children's), or CAN (Free-Standing Cancer) face overlapping regulatory constraints—they cannot purchase outpatient drugs through GPO contracts unless genuinely unavailable elsewhere. Proving that exception requires evidence scattered across multiple systems: purchase records, shortage data, and eligibility records from three separate Bluesight products. No single product held the full picture, and customers repeatedly asked for an AI layer that could reason across boundaries.
The architectural choice to separate AI reasoning from the data layer proved decisive. Rather than expose raw databases to the agent, Bluesight wrapped existing API endpoints in AWS Lambda functions that returned structured, agent-optimized data. This reduced query latency from 5 minutes to 10 seconds—a tenfold improvement that made real-time compliance analysis feasible. The decision also kept business logic in the application layer, preventing the agent from becoming a monolithic decision-maker and maintaining auditability, a non-negotiable requirement under HIPAA.
The compressed timeline—nine months from sprint to production—reflects the maturity of Bedrock's AgentCore Runtime for healthcare workloads. HIPAA eligibility, serverless hosting with session isolation, and built-in MCP (Model Context Protocol) transformation of existing APIs meant Bluesight did not need to build foundational infrastructure. The team deployed in a VPC with encryption, authentication, observability, and infrastructure-as-code already in place, avoiding the architectural pivots and technical debt that typically delay AI projects. A second, more ambitious multi-product version is expected later in 2026.
No comments yet. Be the first to share your thoughts!
Log in to join the discussion
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
1 minute a day. The AI essentials.
200+ sources · Email / LINE / Slack
