Understanding economic perception and financial preparedness from survey responses


Client, a USA-based financial service provider, conducts national surveys online to identify potential consumer segments to offer their products and services. Although, the client was already leveraging the structured responses in the survey to generate insights, nuance-rich open-ended responses in the survey were left unanalysed due to the prohibitive costs. The client wanted to leverage automation to analyse the open-ended responses as they captured rich insights regarding the respondent’s financial planning and confidence and perception about various personal finance and economic factors. To automate the analysis of the open-ended responses and drive insights from them, the client partnered with Inxite Out.


An end-to-end solution was developed leveraging MEGHNAD, Inxite Out’s Conversation Intelligence Framework. The solution consisted of the following steps:


Financial Theme Extraction: End-to-end process using Meghnad

Theme Configuration

Inxite Out collaborated with the client to understand the key business objectives behind the open-ended survey questions and identified 20+ themes across 5+ hierarchical buckets (e.g., Financial Planning, Financial Confidence, Economic Factors, Personal Factors, Employer Perception etc.) to understand the respondent’s confidence in the economy and the state of their financial preparedness. Translated these business requirements into configurations accordingly, for subsequent usage in MEGHNAD.

Financial Theme Extraction

Leveraged MEGHNAD NLU engine, which is built on top of cutting-edge LLMs (large language models), to extract relevant financial preparedness and economic perception themes from the open-ended survey responses of each respondent.


  • Economic Perception & Financial Preparedness understanding obtained for each respondent, as opposed to sample assessment previously – generated richer and more complete insights.
  • Theme understanding achieved at <30% of the estimated cost with manual approaches and at a fraction of the manual analysis time

Case Studies