PwC and AWS Develop Ethical AI Utilizing Automated Reasoning on Amazon Bedrock

This article is a collaborative guest post featuring insights from Scott Likens, Ambuj Gupta, Adam Hood, Chantal Hudson, Priyanka Mukhopadhyay, Deniz Konak Ozturk, and Kevin Paul of PwC

Organizations are increasingly adopting generative AI solutions while ensuring accuracy, security, and compliance. In today’s competitive landscape, speed is gaining prominence over scale, and innovation is paramount, as highlighted in PwC’s recent 2025 business insights on AI agents. To preserve their competitive edge, organizations must facilitate swift deployment and foster trustworthiness in AI outputs. Particularly in regulated sectors, mathematically verifying results can shift the narrative from risk to opportunity in innovation.

This article discusses how AWS and PwC are innovating new reasoning checks that merge extensive industry knowledge with Automated Reasoning checks in Amazon Bedrock Guardrails to drive innovation. Automated Reasoning, a subdivision of AI, focuses on algorithmically searching for mathematical proofs. Automated Reasoning checks in Amazon Bedrock Guardrails employ formal logic to verify if outputs from large language models (LLM) are plausible, and are now publicly available as of August 6, 2025.

This updated guardrail policy ensures accuracy within set boundaries, contrasting traditional probabilistic reasoning methods. The system analyzes AI-generated content against regulations derived from policy documents, inclusive of company policies and operational standards. Automated Reasoning checks yield findings that reveal whether the AI-generated content adheres to the extracted rules, identifies any ambiguities within the content, and suggests ways to eliminate assumptions.

“In a domain where innovations are occurring at remarkable speeds, reasoning is one of the most crucial technological advancements supporting our mutual clients’ success in generative AI,” states Matt Wood, Global CTIO at PwC, during AWS Re:Invent 2024.

Transformative Industry Use Cases Leveraging Amazon Bedrock Automated Reasoning Checks

The strategic partnership merging PwC’s extensive expertise with AWS’s cutting-edge technology is poised to revolutionize how businesses engage with AI-driven innovation. The following diagram illustrates PwC’s implementation of Automated Reasoning, initially aimed at highly regulated sectors like pharmaceuticals, financial services, and energy.

In the sections that follow, we will explore three pioneering use cases developed by PwC teams.

Compliance with the EU AI Act for Financial Services Risk Management

The European Union (EU) AI Act mandates organizations to categorize and verify all AI applications according to particular risk levels and governance criteria. PwC has created a practical framework to tackle this challenge using Automated Reasoning checks in Amazon Bedrock Guardrails, transforming compliance into a systematic, verifiable process rather than a manual chore. Given a description of an AI application’s use case, the solution translates risk classification criteria into defined guardrails, enabling organizations to consistently evaluate and monitor AI applications while aiding expert human judgment through automated compliance verification with auditable artifacts. The primary advantages of utilizing Automated Reasoning checks include:

• Automated categorization of AI use cases into risk groups
• Verifiable logical traces for AI-generated classifications
• Faster identification of necessary governance controls

The following diagram depicts the workflow for this use case.

Review of Pharmaceutical Content

PwC’s Regulated Content Orchestrator (RCO) is a globally scalable, multi-agent capability powered by a specialized rules engine tailored to company, region, product, and usage indications, automating compliance across medical, legal, regulatory, and branding domains. The RCO team was an early collaborator with Amazon Bedrock Automated Reasoning checks, employing it as a secondary validation layer in the marketing content creation process. This reinforced existing content controls, expediting the content creation and review processes while elevating compliance standards. Key benefits of Automated Reasoning Checks within Amazon Bedrock Guardrails include:

• Implements automated, mathematically grounded safeguards for verifying RCO’s assessments
• Facilitates transparent quality assurance with traceable, audit-ready reasoning
• Protects against potentially erroneous or hallucinated outputs

The following diagram illustrates the workflow for this use case.

Real-time Decision Support for Utility Outage Management

Utility outage management leverages Automated Reasoning checks in Amazon Bedrock Guardrails to improve response times and operational efficiency for utility companies. The solution can generate standardized protocols based on regulatory guidelines, create procedures in accordance with NERC and FERC mandates, and validate AI-generated outage classifications. Through an integrated cloud-based framework, this solution implements severity-based verification workflows for dispatch decisions—normal outages (3-hour target) assign tickets to available crews, medium severity (6-hour target) triggers expedited dispatch, and critical incidents (12-hour target) activate emergency protocols with proactive communication.

The key advantages of utilizing Automated Reasoning checks include:

• Improved and enhanced customer responses
• Timely operational insights with verified regulatory compliance
• Accelerated decision-making grounded in mathematical certainty

The following diagram illustrates the workflow for this use case.

Future Outlook

As AI adoption advances, particularly with agentic AI, the AWS and PwC alliance is dedicated to:

• Expanding solutions integrated with Automated Reasoning checks across additional industries
• Creating industry-specific agentic AI solutions featuring built-in compliance verification
• Enhancing explainability features for greater transparency

Conclusion

The integration of Automated Reasoning checks within Amazon Bedrock Guardrails and PwC’s extensive industry expertise presents a robust pathway for deploying AI-based solutions. As a critical element of responsible AI, Automated Reasoning checks establish safeguards that enhance the reliability of AI applications. With a focus on verifiable trust and mathematical certainty in AI outputs, organizations can drive innovation without sacrificing accuracy, security, or compliance. For more information on how Automated Reasoning checks operate, check out Minimize AI hallucinations and achieve up to 99% verification accuracy with Automated Reasoning checks: Now available and Enhance accuracy by incorporating Automated Reasoning checks in Amazon Bedrock Guardrails.

Discover how Automated Reasoning checks within Amazon Bedrock can enhance the reliability of your generative AI applications. To explore the utilization of this feature or to discuss tailored solutions for your specific needs, reach out to your AWS account team or an AWS Solutions Architect. Connect with the PwC team to discover how the combined strength of AWS and PwC can drive innovation in your sector.

---

About the Authors

Nafi Diallo is a Senior Automated Reasoning Architect at Amazon Web Services, where she drives advancements in AI safety and Automated Reasoning systems for generative AI applications. Her expertise includes formal verification methods, AI guardrails implementation, and assisting global customers in building trustworthy and compliant AI solutions at scale. She has a PhD in Computer Science focusing on automated program repair and formal verification and an MS in Financial Mathematics from WPI.

Adewale Akinfaderin is a Sr. Data Scientist–Generative AI at Amazon Bedrock, contributing to innovative developments in foundational models and generative AI applications at AWS. His expertise encompasses reproducible and comprehensive AI/ML methods, practical implementations, and aiding global customers in formulating and developing scalable solutions for interdisciplinary challenges. He holds two graduate degrees in physics and a doctorate in engineering.

Bharathi Srinivasan is a Generative AI Data Scientist at the AWS Worldwide Specialist Organization, focusing on developing Responsible AI solutions, including ensuring algorithmic fairness, validating large language models, and enhancing explainability. Bharathi supports internal teams and AWS clients in navigating their responsible AI journey and has presented her work at various educational conferences.

Dan Spillane, Principal at Amazon Web Services (AWS), spearheads global strategic initiatives within the Consulting Center of Excellence (CCOE). He collaborates with customers and partners to address critical business challenges using innovative technologies, specializing in generative AI and responsible AI, including automated reasoning. His expertise delivers measurable business value at scale. As a passionate learner, Dan studies global cultures and business mechanisms, enhancing his mentorship capabilities and fostering cross-cultural initiatives.

Aartika Sardana Chandras is a Senior Product Marketing Manager for AWS Generative AI solutions, concentrating on Amazon Bedrock. With over 15 years in product marketing, she is dedicated to empowering customers as they navigate the complexities of the AI lifecycle. Aartika is passionate about helping clients utilize powerful AI technologies in an ethical and impactful manner.

Rama Lankalapalli is a Senior Partner Solutions Architect (PSA) at AWS leading a global team of PSAs supporting PwC, a prominent global systems integrator. He champions enterprise cloud adoption by maximally utilizing AWS services across migrations, modernization, security, AI/ML, and analytics. Rama designs scalable solutions that facilitate organizations in accelerating their digital transformation while delivering measurable business results. His leadership integrates technical expertise with strategic insights to foster customer success through innovative, industry-specific cloud solutions.

Scott Likens is the Chief AI Engineer over Global and US teams at PwC, leading the AI Engineering and Emerging Technology R&D teams domestically, driving the firm’s AI, Blockchain, VR, Quantum Computing, and other disruptive technology strategies. With over 30 years of experience in emerging technologies, he has aided clients in transforming customer experience, digital strategies, and operations across various industries.

Ambuj Gupta is a Director in PwC’s AI and Digital Contacts & Service practice, situated in Chicago. With over 15 years of expertise, Ambuj specializes in Artificial Intelligence, Agentic and Generative AI, Digital Contact Solutions, and Cloud Innovation across numerous platforms and sectors. He is acknowledged for driving strategic transformation through Cloud Native AI Automation and emerging technologies—like GenAI-powered agents, Intelligent Agent Assists, and Customer Data Platforms—to boost channel efficiency and employee productivity.

Adam Hood is a Partner and AWS Data and AI Leader at PwC US. As a strategic and results-oriented technology leader, Adam focuses on enterprise-wide transformation and unearthing business value via digital systems, data, and conventional AI/ML, including the development of agentic workflows. With a proven history in industry and consulting, he has guided organizations through intricate digital, financial, and ERP modernization efforts from strategy and business case development to seamless execution and global deployment.

Chantal Hudson is a Manager in PwC UK’s AI and Modelling team. With slightly over five years at PwC, she began her career with the South African firm. Chantal primarily collaborates with major banks on credit risk modeling and has a strong interest in applying AI to enhance modeling practices.

Priyanka Mukhopadhyay is a Manager in PwC’s Cloud and Digital Engineering practice. An AWS Certified Solution Architect – Associate with over 13 years in Data Engineering, she has developed expertise in AWS services and over 12 years of experience in delivering solid projects adhering to Agile methodologies.

Deniz Konak Ozturk is a Senior Manager in PwC’s AI & Modelling team, with around 15 years’ experience in AI/Gen AI and traditional model development, implementation, and validation across UK and EU/non-EU regions, along with compliance assessments with EU regulations and IFRS9 audits. Over the past 6 years, Deniz has primarily focused on AI/Gen AI, showcased through her involvement in developing AI Validation frameworks, implementing this framework for different clients, managing a platform for automated ML, and leading R&D initiatives on Alternative Data Usage for ML-based Risk Models targeting underserved financial segments.

Kevin Paul is a Director in the AI Engineering group at PwC, specializing in Applied AI, with extensive experience throughout the AI lifecycle, building and maintaining solutions across various sectors.