Deana.AI® | Enterprise AI Framework
Deana.AI is a machine learning framework that interprets user inputs in both text and voice, retrieves the most relevant response from its contextual database, and improves with each user interaction. INTERSOG assembled a specialized team comprised of experts in machine learning, natural language understanding, data analysis, and system design to create a robust AI-driven enterprise framework. This team included 12 members: a software architect, 4 full-stack developers, 2 data scientists, 2 database developers, 2 QA engineers, and a project manager. Their project tasks involved: 1) researching NLP frameworks and prototyping applications 2) tuning the NLU pipeline 3) developing an integrated question-answering system 4) generating NLU training data 5) classifying documents for the knowledge base 6) summarizing user inputs 7) implementing optical character recognition (OCR) 8) classifying statistical data 9) generating reports 10) creating a context for the GPT-2 model 11) producing summaries as contextual memory. The first prototype was unveiled two months after the project began, followed by the development of a minimum viable product (MVP) representing Deana in the following months. At launch, Deana.AI utilized text recognition, intent generation, question-answering, and machine learning technologies to automate routine tasks, develop enterprise-wide knowledge bases, and provide decision-making insights for business users. Throughout the development, INTERSOG’s team furthered their experimental research and optimization efforts. Deana.AI® serves as a personal data analyst, offering actionable insights, improving decision-making, and facilitating business growth.