GenAI has emerged as a powerful catalyst, revolutionizing operations across various industries. At GoApptiv, a prominent health tech company, we are actively integrating GenAI into our daily operations and applications. One of the remarkable applications of GenAI lies in streamlining the registration of brands and products. In this article, we illuminate the significant impact GenAI has had on the seamless registration of brands and their corresponding products within our robust product management system.
Streamlining Brand Registration
Efficiently registering all available brands posed a pivotal challenge, particularly given the state of the existing data. The prior master list contained unformatted entries, with some brands lacking organizational mappings. To address this, we harnessed the power of GPT-4, leveraging the OpenAI platform. The results were indeed promising — the AI system efficiently processed the data, providing us with a clean dataset where brands were accurately mapped to their respective organizations.
Enhancing Product Registration
Product registration presented a unique challenge, considering each client submitted product names in their own distinct formats. When devising a standardized approach, two options emerged: either directly inputting the products into the system or adhering to a uniform naming convention for product and variant names, among other fields.
Prior to integrating an ordering system into our applications, we used to directly dump the product names into the system. However, upon implementing the product ordering system, it became imperative to have all product names standardized to ensure consistency in product presentation to our customers.
Before the adoption of GenAI, our team manually created product names, endeavouring to adhere to established standards. Nonetheless, this manual process was susceptible to errors. To enhance efficiency and accuracy, we seamlessly integrated ChatGPT into our workflow.
The GenAI-Powered Prompt
To generate structured and standardized product names, we employed a specific prompt that leveraged GenAI capabilities. The prompt provided clear rules and examples to guide the transformation of input strings into product names and variant names, and it also flagged whether the product was related to the pharmaceutical domain.
We used the following prompt to generate names, the below-mentioned prompt is the minimised version of the original prompt.
Convert these strings to the product name, and variant name and also mark if the product is pharma product or not
Strings:
LUHON CREAM 1 % 30 G x 1
MAMDEW GEL 7.50 % 10 G x 1
Example 1:
String: DIABETONE CAPS x 15
Product Name: Diabetone Capsules
Variant Name: Diabetone Capsules 15's
Example 2:
String: RABETO PLUS CAPS SR 20 MG x 1
Product Name: Rabeto Plus Capsules SR 20mg
Variant Name: Rabeto Plus Capsules SR 20mg
Rules:
1) measurements units like ml, mg or any other will always be in small cases.
2) product quantities will always be mentioned in n's, example x10 will be 10's
3) names will be title case
provide the output in JSON Array format with keys productName, and variantName, isPharmaProduct do not add any comments.
By employing this prompt and continuously refining it with emerging cases, we efficiently processed new data, swiftly generating structured product names. This drastically reduced the time and effort required by our team, ensuring a more efficient and error-free product registration process.
We've developed an automated Python script that seamlessly processes input from Excel and generates the desired output, streamlining the entire workflow.
In conclusion, GenAI stands as a game-changer for GoApptiv in the realms of brand and product registration. Its successful application showcases its potential to streamline operations and enhance data accuracy, solidifying its position as a vital tool for modern businesses striving for efficiency and precision.
This represents just one of the remarkable use cases of GenAI at GoApptiv, and we are continuously exploring more innovative applications. Stay tuned for more insightful use cases in the future. If you found this article valuable, we appreciate your support through likes and shares.
