Skip to content

Latest commit

 

History

History
145 lines (103 loc) · 6.53 KB

n1.md

File metadata and controls

145 lines (103 loc) · 6.53 KB

Use Case 1: GenUI for Code Synthesis

Business Case QW1

Date: May 27, 2024
Submitted by: Mark [Your Last Name]
Project Manager: [Project Manager's Name]

Name Use Case: GenUI for Code Synthesis

Summary Use Case ACTUAL without AI with number of employees:
Currently, our coding team of 10 developers manually writes code for applications. This involves significant time for drafting, reviewing, and correcting code, averaging 60 minutes per application.

Target with application of AI with employee number:
With GenUI, the coding team will use predefined structured prompts for code synthesis, potentially reducing the number of developers required for initial coding stages from 10 to 6.

Data situation - quality, quantity, acquisition, labeling, etc.:
We have a large repository of historical code that is well-documented and labeled, totaling around 1 million lines of code. The data quality is high, with comprehensive comments and structure.

Gainings ACTUAL without AI TARGET with AI

  • per application [min]:

    • Without AI: 60 minutes per application.
    • With AI: 30 minutes per application.

  • repetitiveness per month [n]:

    • Without AI: 100 applications.
    • With AI: 100 applications.

  • per month [min]:

    • Without AI: 60 minutes/application * 100 applications = 6000 minutes.
    • With AI: 30 minutes/application * 100 applications = 3000 minutes.

  • extrapolation per year [h] (taking into account the repetition frequency):

    • Without AI: 6000 minutes/month * 12 months / 60 = 1200 hours.
    • With AI: 3000 minutes/month * 12 months / 60 = 600 hours.

  • time gain [€]:
    Assuming an average developer hourly rate of €50:

    • Without AI: 1200 hours * €50 = €60,000.
    • With AI: 600 hours * €50 = €30,000.
    • Time gain: €60,000 - €30,000 = €30,000.

  • employee savings per year [€] (voluntarily!):
    Assuming we reduce 4 developers:

    • Average annual salary per developer: €70,000.
    • Savings: 4 developers * €70,000 = €280,000.

  • quality gain per year [€] (e.g., reduce error rate, best guess):
    Estimating a 10% reduction in errors, saving about €20,000 annually in debugging and rework.

  • Other gainings per year [€] less hardware, room space (voluntarily!):
    Potential savings in office space and hardware, estimated at €10,000 per year.

  • Savings per year [€]:
    Total Savings: €30,000 (time gain) + €280,000 (employee savings) + €20,000 (quality gain) + €10,000 (other savings) = €340,000.

  • Cost incurred development in h, Hardware, Trainings, etc.:

    • Development: 300 hours * €50/hour = €15,000.
    • Hardware: €5,000.
    • Training: €10,000.
    • Total Cost: €30,000.

  • Risks constraints, assumptions, dependencies, etc.:
    Risks include data privacy concerns, reliance on high-quality training data, and the need for ongoing model updates. Dependencies include stable AI platform availability and continuous data acquisition.

  • Additional opportunities:
    Improved scalability, ability to handle more projects simultaneously, enhanced employee satisfaction.

  • Comments priority, quick win, etc.:
    This project is a high-priority quick win due to its significant cost savings and efficiency improvements.

Use Case 2: Assistance with Code Completion and Debugging

Business Case QW1

Date: May 27, 2024
Submitted by: Mark [Your Last Name]
Project Manager: [Project Manager's Name]

Name Use Case: Assistance with Code Completion and Debugging

Summary Use Case ACTUAL without AI with number of employees:
Currently, our coding team of 10 developers manually writes and debugs code. This process is time-consuming and prone to errors, averaging 40 minutes per debugging session per application.

Target with application of AI with employee number:
Using tools like GitHub Copilot, the coding team will be able to reduce the time required for code completion and debugging, potentially reducing the number of developers needed from 10 to 7.

Data situation - quality, quantity, acquisition, labeling, etc.:
Our existing codebase is extensive and well-documented, providing high-quality data for training and improving AI assistance tools, totaling around 1 million lines of code.

Gainings ACTUAL without AI TARGET with AI

  • per application [min]:

    • Without AI: 40 minutes per debugging session.
    • With AI: 20 minutes per debugging session.

  • repetitiveness per month [n]:

    • Without AI: 150 debugging sessions.
    • With AI: 150 debugging sessions.

  • per month [min]:

    • Without AI: 40 minutes/session * 150 sessions = 6000 minutes.
    • With AI: 20 minutes/session * 150 sessions = 3000 minutes.

  • extrapolation per year [h] (taking into account the repetition frequency):

    • Without AI: 6000 minutes/month * 12 months / 60 = 1200 hours.
    • With AI: 3000 minutes/month * 12 months / 60 = 600 hours.

  • time gain [€]:
    Assuming an average developer hourly rate of €50:

    • Without AI: 1200 hours * €50 = €60,000.
    • With AI: 600 hours * €50 = €30,000.
    • Time gain: €60,000 - €30,000 = €30,000.

  • employee savings per year [€] (voluntarily!):
    Assuming we reduce 3 developers:

    • Average annual salary per developer: €70,000.
    • Savings: 3 developers * €70,000 = €210,000.

  • quality gain per year [€] (e.g., reduce error rate, best guess):
    Estimating a 15% reduction in errors, saving about €25,000 annually in debugging and rework.

  • Other gainings per year [€] less hardware, room space (voluntarily!):
    Potential savings in office space and hardware, estimated at €10,000 per year.

  • Savings per year [€]:
    Total Savings: €30,000 (time gain) + €210,000 (employee savings) + €25,000 (quality gain) + €10,000 (other savings) = €275,000.

  • Cost incurred development in h, Hardware, Trainings, etc.:

    • Development: 200 hours * €50/hour = €10,000.
    • Hardware: €5,000.
    • Training: €8,000.
    • Total Cost: €23,000.

  • Risks constraints, assumptions, dependencies, etc.:
    Risks include data privacy concerns, reliance on high-quality training data, and the need for ongoing model updates. Dependencies include stable AI platform availability and continuous data acquisition.

  • Additional opportunities:
    Enhanced scalability, ability to handle more projects simultaneously, improved employee satisfaction.

  • Comments priority, quick win, etc.:
    This project is a high-priority quick win due to its significant cost savings and efficiency improvements.