15.1 C
New Delhi
Tuesday, November 26, 2024
HomeTechONDC evaluating solutions to add more regional languages

ONDC evaluating solutions to add more regional languages


The government’s ecommerce platform Open Network for Digital Commerce (ONDC) is evaluating better solutions to make more regional languages available on the network.


On September 30, ONDC went live across 16 pin codes in Bengaluru.

“We have been encouraging Indic language support. Enabling Indic language support both for buyer applications and seller applications is something that we’re considering,” T Koshy, chief executive of ONDC, told ET. “There are some solutions that we are encouraging. We have not yet taken a call on our course of action.”

In June, Minister of Commerce and Industry Piyush Goyal had also said ONDC should be made available in regional languages.

Spicemoney and Mystore do not offer their interfaces in regional languages, Koshy said. Even seller apps (seller-facing interfaces) are in English, including their digitised catalogues.

“We (ONDC) are not a funding agency. We will work with multiple agencies like NABARD, SIDBI, and Ministry of MSME etc so that they can develop programmes and schemes to support innovation in this area. This is a part of our mandate to help network expansion,” he added.

Discover the stories of your interest



“As the network expands, we are sure that many more buyer applications (customer-facing interfaces) and seller apps will develop Indic language support. We are trying to find ways to help them in this journey. This is one of our priority areas,” Koshy said.

One of the prize-winning solutions in the ONDC-NABARD hackathon held in July was for digital catalogues, which will help farmers create an online catalogue with ease.

To help farmers sell online, the catalogue uses not only images, description, and quantity, but also text/voice in the needed regional languages.

Another solution was that of search.

The farm input chain is digitized, but there are so many varieties of products/brands that it becomes overwhelming for farmers to choose the right product. A contextual search aims to understand the need and purpose for which a product is being searched and allows farmers to search in their regional languages so that they can understand the products better.

“We were targeting Farmers Producers’ Organizations (FPOs) run by farmers who come together from multiple villages and pull together resources to own a warehouse to store agricultural produce. They are not digitized. Brokers who buy vegetables from farmers and sell it in the mandi communicate over WhatsApp. There is no record of transactions,” Abhishek Padhi, CEO, Subnub Technologies, one of the NABARD-ONDC Hackathon winners, told ET.

Padhi said farmers found it complicated to use the inventory management system. “We thought of a system where just by using QR codes they can mark their inventory,” he said.

The code is embedded with catalogue information like the name of the vegetable, its breed, turnaround time for delivery etc. Stickers with name, size, grade, colour in regional languages can be printed and stuck on the sacks of produce. “We will need funding to implement this solution,” Padhi said.

Thoughtworks has been involved in building two open-source platforms – one of which is called Vakyansh – for speech recognition and translation.

“We have open-sourced the same and made it available to everyone for business-specific customizations and use,” said Rajesh Rajagopalan, industry principal for retail, Thoughtworks India, a winner of the hackathon.

The regional language/Indic language support capability enables back-and-forth translation of text and voice inputs combined with the business context of the domain. This capability has multiple use cases and can be used as a whole or separately in parts.

This means the digital catalogue and search can use the Indic language capabilities as its base components because it is open source.

“The capabilities of the developed solution are limitless. For instance, the showcase at the hackathon focused on a voice-driven application, but the same construct can be directly extended to text and other facets required to support Indic languages,” Rajagopalan added.

Arjun Krishna Shukla, CEO, Cabin4j, another winner of the hackathon, told ET, “We developed an end-to-end algorithm to understand the intent of a search by farmers. This is contextualised search. This will be able to give better results for the search. We have handed over the solution to ONDC. They are free to use it.”



Source link

- Advertisment -

YOU MAY ALSO LIKE..

Our Archieves