Brand sustainability score analysis software

The software autonomously computed sustainability scores with an accuracy exceeding 92%.

Client

E-commerce company in the sustainable apparel business.

Objective

Client was looking to offer environment-friendly apparel products on its platform and allow customers view the detailed sustainability score for each brand. The score had to be calculated by web scraping public data from vendor-neutral websites and analyzing it using the AI/ML engine based on multiple environmental and socially responsible categories.

Solution

Our AI team analyzed massive amounts of data for selected brands, designed robust ML engine based on the GPT model and trained it on the combination of actual and synthetic data. Once AI engine was ready, we built an automated web scraping module and a web-based interface to allow company staff interact with the tool.

The software includes the following capabilities:

Find relevant sustainability-related links on the public internet by using proper keywords
Edit keywords in the application window to setup a more precise custom search
Ability to add custom links manually to the result set
Scrape text from the target links in HTML and PDF formats and display that text along with highlighted excerpts in the application
Run AI model to analyze the scraped text and capture relevant information according to sustainability score rules
Display the results of AI model analysis on a dedicated page
Rank the AI model execution results and calculate the sustainability score for the brand
Display the ranking and scores in the application
Download sustainability score report in PDF format

Result

Our Brand Sustainability Score Analysis Software not only streamlined the evaluation process but also significantly reduced manual efforts, showcasing the transformative impact of AI in promoting sustainable practices in the e-commerce industry.

Automated score calculation: the software automatically calculated sustainability scores with over 92% accuracy.
Efficiency gain: reduced manual analysis efforts by 49%.