Free UiPath-SAIAv1 sample questions

Correct answer: Create a single Group named 'Finance', a single Category named 'Accounts Payable', and two distinct Document Types: 'Invoice' and 'Purchase Order'. Define 'Vendor Name' in both Document Types and 'Payment Terms' only in the 'Invoice' Document Type.

The Taxonomy Manager allows for a hierarchical structure (Group > Category > Document Type). Fields are defined at the Document Type level. Since 'Vendor Name' is needed for both but 'Payment Terms' only for one, defining them as distinct Document Types within the same Category allows for specific field definitions per type while maintaining logical grouping.

Correct answer: Set the 'ApplyOcrOnPdf' property to 'Auto' and use the UiPath Document OCR engine.

The 'Auto' setting determines if the PDF has embedded text. If yes, it extracts it directly; if no (or for images within), it applies OCR. UiPath Document OCR is optimized for document processing workflows.

Correct answer: Intelligent Keyword Classifier

The Intelligent Keyword Classifier classifies documents based on word vectors found within the document. It is effective for distinct document types and allows for learning from a small set of examples during configuration without full ML model training.

Correct answer: Use a Hybrid approach: Configure a Regex Based Extractor for the labeled instances and a Machine Learning Extractor for the unlabeled instances. Prioritize the Regex Extractor in the 'Configure Extractors' wizard.

A hybrid approach leverages the determinism of rules (Regex) for standard cases and the generalization of AI (ML Extractor) for complex, unlabeled cases. Prioritizing Regex ensures speed and accuracy for the 95% standard cases.

Correct answer: The developer passed the 'ExtractionResults' variable (raw output) to the 'Export Extraction Results' activity instead of the 'ValidatedExtractionResults' variable (output from Validation Station).

The 'Present Validation Station' activity outputs a NEW extraction results object containing the human corrections. If the subsequent Export activity uses the original extraction results, the corrections are ignored.

Correct answer: A web-based tool for labeling documents, defining fields, and exporting datasets for model training.

Document Manager is the interface in Automation Cloud/Suite used to prepare training data (labeling) and manage the schema (taxonomy) for ML models.

Correct answer: Validation

The Validation stage (using Present Validation Station or Action Center) is crucial for reviewing low-confidence extractions before data is exported.

Correct answer: Use a Generative LLM-based Extractor initially for zero-shot capabilities, then transition to a specialized ML model trained on the data collected from validation.

Generative Extractors (LLMs) excel at zero-shot extraction on varied layouts (unstructured) with minimal setup. Collecting validated data allows for training a specialized model later, which is often faster and cheaper for high volumes.

Correct answers: Machine Learning Extractor; Data Extraction Scope

The activity itself is obviously required to perform the extraction.

The Machine Learning Extractor must be placed inside a Data Extraction Scope activity to function; it cannot run standalone.

Correct answer: False

Digitization is purely technical (converting file to text/DOM). It does not care about business logic (Taxonomy) yet. Taxonomy is needed for Classification and Extraction.