USE CASES
AI Custom Apps
The Problem
Are Large Language Models or their popular conversational chatbot versions enough for developing your custom AI apps?
- What is more cost-effective for you - developing on your own using large language models or contacting Nowigence with its ready to customize extractive and generative AI techniques, including machine learning and natural language processing?
- Is delivery on-time and on-budget add value to you?
The Situation
When Mr. Cyril-Vincent Michaud, CEO of Crossover Media & Design, asked his team to evaluate AI tools and Large Language Models (LLMs) like ChatGPT, Gemini, Perplexity, and others to find the best fit for their upcoming creative story-telling app called Guppy Stories, they chose the AI toolkit developed by Nowigence.
- Guppy Stories is a vibrant storytelling app that guides children to create their own stories using story “pieces” like characters, locations and situations from a storyworld to create chapters of a new story by creating story chains. Those story “pieces” content are extracted from hundreds of children’s stories and fairy tales connecting them using a proprietary process of relating extracted entities.
- Those story “pieces” content are extracted from hundreds of children’s stories and fairy tales connecting them using a proprietary process of relating extracted entities.
The Solution
The Crossover Media & Design team uploaded many children’s stories during the trial period on the Nowigence platform called researchwork.ai (previously called Lille.ai). Right off the bat, they secured 60 to 95% accuracy (starting accuracy varies depending on the domain) in entity extraction clustered under the right heading.
- Researchwork.ai has an in-built functionality where it is able to cluster entities using an unsupervised model into what the team from Nowigence calls 7Ws (Who, What, Where, When, Why, How, and How Much).
- Selecting a few of the W’s from the drop down menus helps in information retrieval (IR) of the selected few stories.
- IR from the selected few stories is in the form of extracted sentences and ideas.
Nowigence met Crossover’s team custom app specifications on time and on-budget.
