Tyler Owen Morris

A GPT driven chat assistant that uses a Resource Augmented Generative (RAG) approach to give employees instant access to company knowledge bases, and subject matter experts control over the assistant’s responses.

 
 
 

GPT-Driven Slackbots: Revolutionizing Internal Resource Access and Knowledge Sharing

Powered by GPT technology and uniquely augmented with up-to-date expert insights, these bots are transforming the way our company accesses information and shares knowledge.

Three Distinct Bots for Diverse Needs:

  • Qualtrics Bot: Tailored for technical queries about Qualtrics, assisting with coding tasks and offering support for specific technical issues.

  • HR Bot: A resourceful assistant for common human resources queries, streamlining employee access to vital HR information.

  • Advisory Bot: Centrally consolidating advisory resources related to customer experience programs, complementing our Qualtrics consulting services.

Innovative Resource Augmentation:

  • Our bots leverage a Flask-based web interface with a MySQL database accessible to the bots, allowing subject matter experts to input information in a subject-content format, which then becomes immediately accessible to the bots.

  • The bots intelligently categorize user questions using GPT, cross-referencing with our detailed database to provide precise, informed responses.

User Experience:

  • Users interact with the bots through direct messages or channel mentions in Slack.

  • The bots respond with accurate answers, citing the subjects used, enhancing transparency and reliability.

Technical Breakthroughs:

  • Our resource-augmented approach represents a cutting-edge solution in AI, allowing continuous customization without needing to retrain the LLM.

  • GPT-4 Turbo's expanded input tokens significantly enhance scalability and depth of response.

Impact and Benefits:

  • In the initial month, our bots saved an average of 8 hours per week, a figure expected to rise with wider adoption and knowledge base expansion.

  • These tools serve as efficient internal aids, empowering our consultants with quick access to updated technical knowledge.

Security and Data Privacy:

  • We maintain strict HTTPS protocols for data transmission.

  • While user discretion is advised regarding sensitive data, the primary focus is on general information, mitigating major privacy concerns.

Conclusion: Our GPT-driven Slackbots represent a leap forward in internal resource management and knowledge dissemination. By harnessing the power of AI with expert-driven data, we've created a dynamic, efficient, and continually evolving knowledge ecosystem. These bots not only save valuable time but also ensure that our team has the most current and accurate information at their fingertips, fostering a more informed, efficient, and proactive workforce.

Film Predicting Robot

Using real world data from IMDB I trained a Machine Learning Algorithm to predict the opening weekend box office revenue of films. The web application gives users the option to generate random films from all available data-points, from a small list of top actors, directors, and studios, or users can input their own custom films down to the month of release and the rating of the film. You can even input actors and directors that never existed, and the model will still attempt to make a prediction! The app is available at filmpredictingrobot.com

  • Python / Sklearn

  • Flask