Scribd
Upload
43 views2 pages

Problem Statement - Fpgrid

The document describes four challenges for building AI solutions: 1. A smart bag that predicts future grocery orders based on past orders and creates customized order baskets. 2. A conversational credit application interface using chat instead of forms, adapting to product metadata. 3. Identifying social media signals of trust and affluence to assess creditworthiness for those without credit histories. 4. An interactive voice bot for customer engagement at scale with human-like speech synthesis and understanding, and domain-specific natural language processing.

Uploaded by

rajvi yadav
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
43 views2 pages

Problem Statement - Fpgrid

The document describes four challenges for building AI solutions: 1. A smart bag that predicts future grocery orders based on past orders and creates customized order baskets. 2. A conversational credit application interface using chat instead of forms, adapting to product metadata. 3. Identifying social media signals of trust and affluence to assess creditworthiness for those without credit histories. 4. An interactive voice bot for customer engagement at scale with human-like speech synthesis and understanding, and domain-specific natural language processing.

Uploaded by

rajvi yadav
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

1.

Smart Bag Creator Challenge for Flipkart Grocery

Forgot to order fresh coriander or your favorite box of strawberries or tomatoes while placing
your weekly grocery order online? Wondering that you’ll need to place another small order for
the items missed? Well, that’s something all of us have faced while ordering online, and that’s
what this challenge attempts to solve! In this challenge, teams are required to create a smart
bag that, based on past order history, can predict future orders and create a smart order basket
for the customer - selecting the items required in advance and enabling a faster buying
experience.

2. Conversational Credit Application

As we go deeper into customer segments, conversational interfaces could be more engaging


than typical form-based ones. As part of this challenge, teams are expected to build a chat
based conversational interface based on application metadata.
- application metadata will include: field info (label, data type, validation constraints, value
options if applicable, etc.)
- chat interface should adapt to the product-specific form definition(the metadata format will be
preserved)
Preferred tech: React widgets that are themable
Bonus: innovation within the chat experience

3. Trust and affluence signal extraction from social media data

Less than 10% Indians have a credit bureau presence, so it is essential to identify and utilize
alternate forms of data to assess customers’s credit worthiness. As part of this challenge,
teams are expected to identify markers from social media data that indicate trust and affluence;
Generate sub-indicators for each of those markers & subsequently, attempt to utilize all signals
available (ex. posts, sessions, check-ins, social graph, media content, etc.) & come up with an
explainable trust and affluence score/classification based on all the signals
Preferred tech: Open source
Bonus: Signal extraction from multiple social media channels (ex. FB, Instagram, Twitter, etc.)
4. Interactive voice bot for user engagement

Key problem statement: How can we instantly and effectively engage with millions of customers
at zero marginal cost. Teams are required to build an interactive voice bot that is expected to
have human-grade voice synthesis and voice understanding. It should have NLP-based
understanding of context and user intent (trained on domain-specific data) & have ability to
understand mixed-language speech.
- Preferred tech: Open source
- Bonus: Multi-lingual bot (english, hindi), Fintech-specific domain training

You might also like