Consultancy PwC and Amazon Web Services (AWS), announced an expansion of their strategic alliance to accelerate generative AI (GenAI) adoption across industries, including pharmaceuticals, financial services, and energy.
The collaboration with AWS aims to create GenAI-powered applications tailored to specific industry needs.
The integration of Amazon Bedrock’s Automated Reasoning checks helps make sure that these applications are reliable and compliant, preventing factual errors and enhancing decision-making.
“By combining our expertise with AWS’ AI tools, we aim to help businesses transform their operations,” said Paige Hayes, PwC’s global AWS alliance leader.
She explained the collaboration agreement provides PwC with early access to new services being introduced by AWS like Nova and Bedrock.
“We build these and other AWS services into unique industry solutions – many of which are relevant for highly regulated sectors,” she said.
Where they see the most AI usage is from organizations in those highly regulated sectors, including financial services, health care and insurance.
“Companies in those sectors have a strong understanding of the regulations in place, allowing them to accelerate and scale their AI adoption,” she said.
Hayes added the regulatory environment has driven the right kind of investments when it comes to responsible AI usage and data, including access, quality and governance.
Bedrock Automated Reasoning checks play a central role in enabling broader acceptance of GenAI solutions by confirming the validity of GenAI output.
The service prevents factual errors due to hallucinations through checks that use logically accurate and verifiable reasoning.
The aim is to provide customers with confidence in deploying powerful and explainable GenAI solutions.
“At PwC, we integrate these features into existing AWS-based solutions that we have built,” Hayes said.
PwC’s GenAI-powered solutions, including the AI-driven annotation (AIDA) solution are designed to enhance productivity and decision-making for clients across diverse sectors.
The AIDA solution transforms contracts into digital intelligence, both on demand and at scale.
“Using the latest AWS LLMs, companies can quickly extract terms and provisions from a wide variety of unstructured data and agreement types securely and with confidence,” Hayes said.
The solution includes a GenAI chat feature that can help users answer questions about the documents they’re navigating.
PwC is also incorporating the Automated Reasoning checks within industry solutions like the Maestro Medical Legal Review solution.
This integration helps validate that content generated by large language models (LLMs) complies with regulations.
Additionally, these checks are used in digital call centers to help agents adhere to company protocols, thereby enhancing the reliability and compliance of GenAI applications across various industries.
By leveraging Bedrock’s security capabilities and automated reasoning safeguards, Hayes said PwC can help maintain the security, scalability and trustworthiness of GenAI applications for their enterprise clients.
“These features are fundamental in preventing factual errors due to hallucinations, using logically accurate and verifiable reasoning,” she said. “Ensuring we take a comprehensive approach that includes hardening through implementation and use of this technology is also imperative.”
PwC achieves this by performing vulnerability assessments, penetration testing to exam resilience of the system, data encryption analysis, appropriate access control configuration, monitoring and logging of the system.
The company also works to ensure there is a data loss prevention tool employed, which includes the new technology in its scope.
“This approach provides clients with the confidence to deploy powerful and explainable GenAI solutions,” Hayes said.
