Regulatory reporting · Production

Gardenia Technologies helps customers create ESG disclosure reports 75% faster using agentic generative AI on Amazon Bedrock

The problem

ESG reporting has become a significant operational burden: 55% of sustainability leaders cite excessive administrative work in report preparation, and 70% indicate that reporting demands inhibit their ability to execute strategic initiatives. Organizations must collect thousands of data points from multiple sources across multiple evolving disclosure frameworks, and much of the work is highly manual, leaving sustainability teams spending more time managing data collection than developing impactful strategies.

Workflow diagram · grounded in source
1
Agent setup and access configuration
trigger
“The Report GenAI agent is configured and authorized to access an ESG and emissions database, client document stores (emails, previous reports, data sheets), and document searches over the public internet”
2
AI batch-fill of report questions
ai_action
“The agent then iterates through each question and data point to be disclosed and then retrieves relevant data from the client document stores and document searches. This information is processed to produce a response in the expected form…”
3
Cited-source audit review
validation
“Each response includes cited sources and—if the response is quantitative—calculation methodology. This enables users to maintain a clear audit trail and verify the accuracy of batch-filled responses quickly.”
4
Human-in-the-loop editing
human_review
“our approach allows for a human-in-the-loop to review, validate, and iterate on batch-filled facts and figures. In the following figure, we show how users can chat with the AI assistant to request updates or manually refine responses. Wh…”
5
LLM judge quality validation
validation
“Report GenAI uses state-of-the-art LLMs on Amazon Bedrock as LLM judges”
6
Multi-framework repeat and document store update
output
“Users can batch-fill multiple reporting frameworks to simplify and expand their ESG disclosure scope while avoiding extra effort to manually complete multiple questionnaires. After a report has been completed, it can then be added to the…”
7
Expert review feedback loop
feedback_loop
“These expert reviews provide valuable training data that can be used to enhance system performance through refinements to RAG implementations, agent prompts, or underlying language models.”
Reported outcome

Omni Helicopters International cut their CDP reporting time by 75%, completing their 2024 CDP submission in one week instead of the one month it had previously required.

Reported metrics
ESG reporting time reduction75%
OHI CDP reporting time beforeone month
OHI CDP reporting time afterone week
Sustainability leaders citing excessive admin work in report preparation55%
Show all 5 reported metrics
ESG reporting time reduction75%
OHI CDP reporting time beforeone month
OHI CDP reporting time afterone week
sustainability leaders citing excessive admin work in report preparation55%
sustainability leaders whose strategic work is inhibited by reporting demands70%
Reported stack
Amazon BedrockReport GenAIClaude Sonnet 3.5Haiku 3.5LangChainAWS LambdaAmazon ECSAmazon CognitoAmazon Step FunctionsAmazon TextractFaissStreamlitRAGtext-to-SQLPydanticDynamo DBAmazon Titan Text EmbeddingsLance DBAmazon NovaReAct
Source
https://aws.amazon.com/blogs/machine-learning/how-gardenia-technologies-helps-customers-create-esg-disclosure-reports-75-faster-using-agentic-generative-ai-on-amazon-bedrock?tag=soumet-20
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Omni Helicopters International cut their CDP reporting time by 75%, completing their 2024 CDP submission in one week instead of the one month it had previously required.

What tools did this team use?

Amazon Bedrock, Report GenAI, Claude Sonnet 3.5, Haiku 3.5, LangChain, AWS Lambda, Amazon ECS, Amazon Cognito, Amazon Step Functions, Amazon Textract.

What results were reported?

ESG reporting time reduction: 75%; OHI CDP reporting time before: one month; OHI CDP reporting time after: one week; Sustainability leaders citing excessive admin work in report preparation: 55% (source-reported, not independently verified).

How is this regulatory reporting AI workflow structured?

Agent setup and access configuration → AI batch-fill of report questions → Cited-source audit review → Human-in-the-loop editing → LLM judge quality validation → Multi-framework repeat and document store update → Expert review feedback loop.