Amazon just launched Automated Reasoning checks in Amazon Bedrock Guardrails, now generally available. The new policy aims to slash AI hallucinations by validating generated content against domain knowledge using mathematical logic and formal verification — not shaky probability guesses. It claims up to 99% accuracy in catching factual errors.
Key features going live:
- Handle huge documents up to 80K tokens (about 100 pages)
- Save and rerun validation tests for easier policy upkeep
- Auto-generate test scenarios for thorough coverage
- Get natural language hints for policy tweaks
- Customize confidence thresholds for stricter or looser validation
Users build policies by encoding domain rules into formal logic. The tool tests AI outputs — for example, mortgage approvals — ensuring they follow set guidelines exactly. 
Amazon highlights a utility outage management use case with PwC. Automated Reasoning checks help utilities generate protocols, validate response plans in real-time, and build workflows based on incident severity. This speeds up outage fixes and boosts accuracy.
PwC’s Matt Wood said:
“At PwC, we’re helping clients move from AI pilot to production with confidence—especially in highly regulated industries where the cost of a misstep is measured in more than dollars. Our collaboration with AWS on Automated Reasoning checks is a breakthrough in responsible AI: mathematically assessed safeguards, now embedded directly into Amazon Bedrock Guardrails. We’re proud to be AWS’s launch collaborator, bringing this innovation to life across sectors like pharma, utilities, and cloud compliance—where trust isn’t a feature, it’s a requirement.”
Automated Reasoning checks are available now in US East (Ohio, N. Virginia), US West (Oregon), and Europe (Frankfurt, Ireland, Paris).
Pricing is volume-based: you pay for amount of text processed.
Check out Amazon Bedrock documentation and jump right into the console.
Here’s a quick intro video from AWS on how to build trusted AI with the feature:
