How Gardenia Technologies Accelerates ESG Disclosure Reporting by 75% with Generative AI on Amazon Bedrock

Learn About Amazon VGT2 Learning Manager Chanci Turner

As organizations around the globe embark on their sustainability journeys, the adage “What gets measured gets managed” has become a cornerstone principle. Companies are setting benchmarks to monitor their progress, aided by an evolving set of reporting standards, both mandatory and voluntary. However, the task of ESG reporting has transformed into a daunting operational challenge. A recent survey found that 55% of sustainability leaders report excessive administrative burdens in preparing their reports, while 70% say that reporting demands hinder their ability to implement strategic initiatives. This scenario opens the door for generative AI to streamline routine reporting tasks, enabling organizations to reallocate resources towards more impactful ESG initiatives.

Gardenia Technologies, a data analytics firm, has collaborated with the AWS Prototyping and Cloud Engineering (PACE) team to launch Report GenAI, a fully automated ESG reporting solution driven by cutting-edge generative AI models on Amazon Bedrock. This article explores the technology underpinning an agentic search solution that employs Retrieval Augmented Generation (RAG) and text-to-SQL capabilities, assisting customers in cutting down ESG reporting times by as much as 75%.

Understanding the Challenge: The Increasing Complexity of ESG Reporting Requirements

Sustainability disclosures have become integral to corporate reporting, with 96% of the largest 250 companies reflecting on their sustainability efforts based on government and regulatory guidelines. To satisfy reporting mandates, organizations must navigate numerous data collection and process-related hurdles. Compiling a single report entails gathering thousands of data points from various sources, including official documents, databases, unstructured document stores, utility bills, and emails. For instance, the EU Corporate Sustainability Reporting Directive (CSRD) framework requires the collection of 1,200 distinct data points across an organization. Even voluntary disclosures, like the CDP, which includes about 150 questions, span a wide array of topics such as climate risk and water stewardship. Collecting this information can be a long-drawn-out process.

Additionally, many organizations with established ESG programs face the challenge of reporting across multiple frameworks, including SASB, GRI, and TCFD, each with its own reporting and disclosure standards. The ever-evolving nature of reporting requirements often leaves organizations scrambling to stay current. Currently, much of the reporting work is manual, resulting in sustainability teams spending excessive time on data collection and questionnaire responses rather than focusing on developing effective sustainability strategies.

Solution Overview: Streamlining the Heavy Lifting with AI Agents

Gardenia’s innovative approach to enhancing ESG data collection for enterprises is embodied in Report GenAI, an agentic framework utilizing generative AI models on Amazon Bedrock to automate substantial portions of the ESG reporting process. Report GenAI pre-fills reports by leveraging existing databases and document stores, as well as conducting web searches. The agent collaborates with ESG professionals to review and refine responses. This workflow consists of five key steps to automate ESG data collection and assist in curating responses:

1. Setup: The Report GenAI agent is configured to access ESG and emissions databases, client document stores (emails, past reports, data sheets), and perform document searches on the internet. Client data is securely stored within designated AWS Regions using encrypted Amazon Simple Storage Service (Amazon S3) buckets with VPC endpoints for secure access. The agent also accesses relevant ESG disclosure questionnaires containing questions and expected response formats (referred to as report specifications).
2. Batch-fill: The agent processes each question and data point, retrieving relevant data from client document stores and searches. The information is then formatted to meet reporting requirements.
3. Review: Each response includes cited sources and calculation methodologies for quantitative responses, maintaining an audit trail and enabling quick verification of accuracy.
4. Edit: Although the workflow is automated, human oversight is incorporated to review, validate, and refine the batch-filled information. Users can interact with the AI assistant to request updates or modifications, ensuring the final responses meet expectations.
5. Repeat: Users can batch-fill across various reporting frameworks to simplify and broaden their ESG disclosure scope without the added effort of completing multiple questionnaires manually. Completed reports can be archived for future reference, and Report GenAI supports bringing in custom reporting specifications.

Now that we’ve covered the Report GenAI workflow, let’s delve into its architecture.

Architecture Deep-Dive: A Serverless Generative AI Agent

The architecture of Report GenAI consists of six components: a user interface (UI), a generative AI executor, a web search endpoint, a text-to-SQL tool, the RAG tool, and an embedding generation pipeline. These components work together to orchestrate the workflow and generate responses through actions such as web searches and SQL execution. This architecture ensures a robust solution that meets the evolving needs of ESG reporting.

For further insights on workplace challenges, check out this webinar. You may also find valuable strategies for hybrid work environments in this resource. If you’re interested in interview preparation, this link provides excellent guidance.