A company is considering where to apply generative AI (foundation models) within its own services, while distinguishing these cases from situations where traditional approaches are sufficient. Which TWO of the following are use cases where generative AI (foundation models) can be effectively applied? (Choose TWO.)

1 / 1
Select all that apply
CorrectA, E

Explanation

Question Overview

Select two use cases where generative AI can be effectively applied.

Requirements to satisfy
  • 1「use cases where generative AI (foundation models) can be effectively applied」Use cases requiring semantic understanding (search) or content generation
  • 2「Choose TWO」Semantic search and recommendation text generation are applicable
Per-option explanation
ACorrect

Enhancing semantic-based search for internal documents

This is correct. Using embeddings (converting text meaning into numerical vectors) created by foundation models, or RAG, enables meaning-based search beyond simple keyword matching, enhancing the search experience. While it is not strictly 'generating text,' it is a use case that leverages a foundation model's semantic understanding.

BIncorrect

Predicting next month's sales from numerical data using a regression model

This is incorrect. Numerical sales prediction is sufficiently achieved with regression (traditional predictive ML), and is not a use case for foundation models. While it is a valid ML use case, it is not a situation where generative AI (foundation models) is applied.

CIncorrect

Determining inventory reorder points using rule-based conditional branching

This is incorrect. Threshold-based reorder point determination is a process that can be reliably achieved with rule-based mechanisms. There is no need to use generative AI for situations where deterministic rules are sufficient.

DIncorrect

Aggregating access logs and displaying them in a dashboard

This is incorrect. Log aggregation and visualization is a routine process achieved with BI tools or aggregation queries. While it is useful for data utilization, it is not a use case that requires generative AI (foundation models).

ECorrect

Generating personalized product recommendation text tailored to user preferences

This is correct. Using generative AI, personalized recommendation text can be generated tailored to each user's preferences and situation, improving the quality of suggestions.

Key Takeaway

Foundation models can be effectively applied in situations requiring 'semantic understanding (embeddings/search)' and 'content generation (text/dialogue).' On the other hand, for numerical prediction (regression), threshold determination (rule-based), and aggregation (BI), traditional approaches are sufficient. The criterion is 'does this situation require generation or semantic understanding?'

Related Links