top of page

Greg Robison

Sep 160 min read

AI Augmented Sample: Case Study

AI Augmented Sample: Case Study

Our case study explores how synthetic data, created by QuantifAI and large language models (LLMs), can be used alongside real human data to yield results comparable to much larger human sample sizes

Using data from 1,500 human respondents, there’s no need for synthetic augmentation, making this large sample a strong baseline for comparison. Two experiments assess the accuracy and reliability of results with reduced human samples of 300 and 150.

1,500 SYNTHETIC RESPONDENTS BASED ON 300 HUMAN RESONDENTS | CORRELATION = 0.95 (95%)

1,500 SYNTHETIC RESPONDENTS BASED ON 159 HUMAN RESONDENTS | CORRELATION = 0.90 (90%)

We can reduce the need for extensive human data collection, particularly for hard-to-reach populations, through augmentation by generating accurate synthetic data based on a small, high-quality human sample.

Selected clients including Microsoft, T-Mobile, Amazon, Meta, Randstad

button that says "let's stay connected"

Recent Posts

What is ChatGPTs Favorite Flavor of Ice Cream?

5 Ways F'inn is Using Generative AI

Specialized Intelligence

bottom of page