Feature Reclaimation

Challenge:
Recover valuable features that were lost during development.

Solution:
Create a way to collect features that had been removed from development and reintroduce them.

Features

  • Recovered Features

  • Identify Urgent Features

  • Jira Compatible

  • DevOps Compatible

AI Models

  • RAG Pipeline

  • Gradient Boosted Trees

  • Large Language Model

User Experience

  • Sharepoint Dashboard

  • Sharepoint Listing

  • Journey Map

  • ML Pipeline

Sharepoint Templates

Challenge

Sharepoint 3rd Party Dashboard
A web-based dashboard will show how trends are seen across time.

 

Sharepoint 3rd Party List
A web-based dashboard will show how trends are seen across time.

User Journey

The journey map shows the most critical peer pathways.

The purpose is to reach a cohesive truth, and the journey map is a way to show the level of consistency within a product experience. It helps us understand the product on a high level while informing us where we can improve in the future product roadmap.

Also, it is an excellent way for user experience intention to continue to encourage understanding among the key stakeholders. It can also add clarity to business.

Agentic AI

  • The Historian

    Retrieval Augmented Generation
    (e5-large-v2 + SaaS GPT4)

    The historian searches through old agil product management software to find features that have been lost in development.

  • The Auditor

    Gradient Boosted Trees
    (XGBoost1)

    This model can analyze and rank features based on value and feasibility. Ranking can express a level of urgency as well as potential ROI.

  • The Conductor

    Large Language Model
    (SaaS GPT4)

    The composer can take a look at the highest ranked features, and provide an explanation for them as well as explain why the feature is important to the customer base and sales team.

MIT xPRO
AI in Healthcare Fundamentals & Applications

GluCozie was initially the thesis for the program. Since then, it has taken on a life of its own and has been progressing into the commercial space. The design and AI work shown here were necessary for a product to prove itself in the market.

The need for an enterprise-class user experience is valuable. This was also an accurate AI solution pipeline that was enhanced by the relevant training.

Built on Existing Models

As part of the requirements, we were tasked with leveraging solutions through a preexisting AI platform. This served to be a boon since much of the question of the AI’s potential was answered already, and the focus was on the innovation in the service.

A New Market

Originally, conceived as a clinical device, a challenge revealed itself regarding the limitations in the AI's ability to detect glucose levels. Specifically, the error range was too great to attain compliance in the US market. Hence, the business development decision is to seek post-clinical and pre-diabetic individuals to be aware of how their personal glucose levels are trending.

The future of GluCozie is nuanced since much of the design work could be suited as a component to existing bioscience products.

Pipeline validated by Microsoft Copilot (GPT-5). Hot Air Balloon FPO grahic was rendered by Microsoft Copilot (GPT-4).
Elder woman was generated by FireFly 3

 

Features

  • Wireless

  • Glucose Trend Detection

  • Non-Clinical Users

AI Models

  • Convolutional Neural Network

  • Variance Auto-Encoder

User Experience

  • Customer Journey

  • Mobile App

  • Web Dashboard