Editja

Ixxerja permezz ta’


Transform unstructured text files into Delta tables by using AI-powered tools (preview)

Use AI-powered shortcut transformations to extract insights from unstructured text files. AI-powered transformations apply language processing to .txt files to summarize content, detect sentiment, translate languages, redact PII, or extract named entities.

Important

Shortcut transformations are currently in public preview. Features and behavior might change before general availability.

Why use AI-powered shortcut transformations?

Shortcut transformations in Microsoft Fabric include a set of built-in, AI-powered transformations that you can apply directly to .txt files referenced through shortcuts. The engine automatically keeps the output Delta table in sync with the source files.

Benefit What it means for you
Accelerate time-to-insight Go from raw text to a queryable Delta table in minutes, no ETL required.
Lower maintenance The transformation engine watches the source folder on a 2-minute schedule, so outputs stay up to date automatically.
Enterprise-grade security PII detection helps you redact sensitive data before it lands in analytics.
Consistent, repeatable results Built-in AI models provide standardized sentiment scores, entity tags, and translations, eliminating manual data-prep drift.

Supported AI transformations

Transformation Purpose
Summarization Generate concise summaries from long-form text.
Translation Translate text between supported languages.
Sentiment analysis Label text sentiment as positive, negative, or neutral.
PII detection Find and redact personally identifiable information (names, phone numbers, emails).
Name recognition Extract named entities such as people, organizations, or locations.

For example, customer feedback stored in a data lake might contain sensitive details like names, emails, and phone numbers. Apply the PII detection transformation to scan and redact this content automatically and produce a privacy-compliant Delta table for analysis.

Note

AI transformations support .txt files only as input.

How it works

  1. Create a shortcut
    Reference a folder of .txt files in Azure Data Lake, Amazon S3, or another OneLake shortcuts source.
  2. Select an AI-powered transformation
    Pick one of the supported transformations during shortcut creation.
  3. Automatic sync
    The engine checks the source folder every two minutes. New, modified, or deleted files are reflected in the Delta table.
  4. Query-ready output
    Use the resulting table immediately in reports, notebooks, or downstream pipelines.

Regional availability

AI-powered transforms are currently available in these regions: Azure Language in Foundry Tools regional support