Local large language models (LLMs) can support workflows that process documents in batches, use a local dataset, need to work without an internet connection, or should keep prompts and source material within a controlled environment. Whether a local model is suitable depends on the task, the model's capabilities, and how it is deployed. Running a model locally does not by itself guarantee privacy or security.

Here are some common use cases:

• Text summarization and drafting:

  Summarizing documents, preparing reports, drafting emails or messages, and generating other text from material you provide.

• Data extraction and classification:

  Finding and classifying information in documents, reports, records, scanned material, emails, and other unstructured data, where the model and processing setup support the input format.

• Knowledge assistants and chatbots:

  Answering questions using documents or data you supply, helping employees find internal information, and supporting service enquiries.

• Knowledge base support:

  Organizing and maintaining internal knowledge resources so employees can find relevant information.

• Personalized content:

  Adapting content to user preferences and needs using information provided for the task.

• Sentiment and text analysis:

  Analyzing the tone of emails, feedback, reviews, or other text to help identify themes and sentiment.

• Offline processing and translation:

  Processing a local dataset or translating text and documents without an internet connection, when the required model and data are available locally.

• Code generation and review:

  Assisting with code completion, automated test-case generation, bug detection, code analysis and review, and documentation. Results should be checked as part of the development process.