Shinydocs Pro Summary

Shinydocs: Advanced Data Extraction (ShinyAI)

Our advanced data extraction AI model is incredibly unique and new-to-market evolving data extraction technology.

Custom ML/AI model applied and tuned for data extraction from flat drawings that can be deployed securely with no data leaks.

My Role:

  • Facilitate internal innovation to a product roadmap
  • Management and alignment of development resources
  • Ran Internal tested and development
  • Creation of a launch plan
  • Creation of short and long term roadmaps
  • Creation and management of product backlog
  • Product integration plan for all Shinydocs products
  • Competitive research
  • Client feedback & interest facilitation
  • Creation of marketing resources and plan
  • RFP/contract response and revision
  • Sales strategy alignment
  • Sold to an early-adopter to prove the value and isolate the biggest wins + biggest needs
  • Actively in development today
  • A major part of the new Shinydocs packages and company vision

Value:

  • Runs and stored locally – no data security breaches or sharing with LLMs (large learning models)
  • Deep Extraction – page-level allows for comprehensive understanding versus document-level
  • Custom to each organization – learn each organizaion’s preferences and data specifically
  • Extraction + Classification – beyond text extraction the model also classifies data

Metrics:

  • Number of pages processed per client per month
    • Northstar Metric
    • Represents overall usage and value to client
    • Higher the better
  • Pages reviewed with changes
    • Quality Metric
    • Represents the need for users to edit/contradict the information generated by the model/product
    • Lower the better – used to plan changes to the model architecture
  • # of metadata attributes per page vs # of pages
    • Volumetric Metric
    • Comparison of data depth vs data breadth
    • Used to determine how to evolve the model to faster and more efficient
  • Time from document ingestion to final migration
    • User Experience Metric
    • Lower the better
    • Used to understand the need for human involvement + system performance

Most credit on this innovation is due for/to Mervin Bowman – the original proprietor of the core technology.

An NDA is in place keeping me from sharing many technical details at this time – contact me to learn more.

Patents are submitted but pending!