Johnson Lambert's AI Transformation with Cohere and Provectus

Overview

Johnson Lambert LLP, a CPA and consulting firm with over 35 years of experience, identified the potential of generative AI to boost productivity, streamline processes, and enhance client services. Partnering with Provectus and Cohere, they implemented an AI-driven solution transforming their audit process.

David Fuge, CIO at Johnson Lambert LLP, highlights: “With the AI solution designed by Provectus and powered by Cohere, Johnson Lambert achieved a 50% reduction in audit hours, allowing our team to focus on high-impact solutions for clients."

The Challenge

Johnson Lambert saw an opportunity to optimize their audit processes for efficiency and scalability. The manual approach to auditing was labor-intensive, requiring 60 to 80 hours per audit, which constrained their ability to engage in more strategic, customer-focused activities. This traditional method was not only time-consuming but also increased the risk of errors, limiting their capacity to scale and improve client services.

Additionally, the technical challenge of processing and validating vast amounts of unstructured data from diverse financial documents further complicated the audit process. To address these complexities, Johnson Lambert needed an advanced AI solution capable of automating and streamlining their workflows, which led them to choose Cohere's state-of-the-art generative AI technology.

The Solution

Johnson Lambert sought a transformative solution to automate and streamline their audit processes to enhance productivity and accuracy. Supported by Provectus, Johnson Lambert implemented a sophisticated AI-powered solution leveraging Cohere Command on Amazon Bedrock.

The solution automates report processing, extracting and validating financial insights from various unstructured PDF documents. Key highlights include:

1. Automation of Data Extraction: Chosen for their cost, performance, and output quality, Cohere Command is trained to recognize and extract relevant financial information accurately from diverse document formats. Cohere models ensure higher quality of information extraction from unstructured PDF files (compared to OCR) and more efficient reference resolution.
   
2. Data Validation: Using a rigorous validation methodology including rules-based checks, anomaly detection, cross-referencing, and human review, was crucial in maintaining the high standards of accuracy and reliability expected in Johnson Lambert's audit reports.
   
3. User-Friendly Interface: A customized interface was developed to allow auditors to interact with the AI, review extracted data, and make necessary adjustments seamlessly.
   
4. Rapid Prototyping and Deployment: The solution was developed and deployed in less than two months, showcasing the agility and effectiveness of both Provectus approach and Cohere technology deployed on Amazon Bedrock.

“The agility and performance of Cohere technology, demonstrated by the rapid prototyping and deployment within two months, was instrumental in helping us iterate and build the best solution,” David Fuge continues.

The Impact

This implementation resulted in a 50% reduction in audit time and a 20% increase in efficiency, empowering Johnson Lambert to focus on more strategic, client-facing activities.

“Cohere Command was key to this success, providing robust capabilities in handling unstructured data with high accuracy,” says David Fuge, CIO at Johnson Lambert. “We are confident this tool will continue to drive efficiency, potentially reaching up to 80% time savings on document analysis, while maintaining the security and integrity of our data within the AWS ecosystem."

Johnson Lambert's success story illustrates how AI is transforming traditional processes, significantly increasing efficiency and productivity. By partnering with Cohere, businesses can harness the power of generative AI to streamline operations and focus on strategic growth.