askEdgi Prompts
Discovery and Analytics Mode
This article outlines the supported prompt patterns, applicable modules, and validated outcomes for the AskEdgi Discovery and Analytics Modes. It also specifies the current functional boundaries and limitations to help in understanding the types of queries effectively supported by the system.
Discovery Mode in AskEdgi is designed to fetch metadata information about data objects and glossary terms. It enables users to identify and explore datasets, schemas, tables, files, reports, and related attributes that can later be analyzed in detail within the workspace. Analytics Mode, on the other hand, focuses on assisting with analytical insights derived from available data.
Supported Object Types
Schema
Retrieves schema-level metadata, including ownership, popularity score, and stewardship details.
Table
Fetches table-level information, including certification, DQ score, row count, and metadata curation score.
Table Column
Provides detailed column-level metadata and relationships to parent tables.
File
Supports retrieving file-level details, including business description, metadata curation, access permissions, and profiling status.
File Column
Retrieves metadata of columns within files and identifies file associations.
Report
Fetches metadata information for registered reports.
Report Column
Retrieves metadata for report columns, including hierarchy and relationships.
API and API Attributes
Retrieves details of APIs, including endpoints, attributes, and metadata references.
Terms and Tags
Supports term-level metadata, stewardship information, and category-based term retrieval.
Projects
Identifies project-level associations and ownership details.
Discovery Mode Capabilities
Discovery Mode supports comprehensive retrieval of metadata attributes, ownership details, governance assignments, associated terms, tags, and asset-level scores across cataloged objects. It enables efficient exploration of schemas, tables, files, reports, APIs, and glossary terms through natural-language prompts, allowing users to quickly access key context without navigating multiple catalog screens.
This capability is optimized for object-level discovery and relationship awareness, ensuring users can locate relevant assets, understand their business and technical attributes, identify stewardship responsibilities, and evaluate metadata quality indicators. By centralizing visibility into catalog metadata, Discovery Mode helps accelerate data understanding, improves trust in enterprise assets, and provides a foundational entry point for further analysis and curation activities.
Prompts
The following categories represent the types of questions successfully processed and validated in Discovery Mode. Each prompt type was tested with relevant objects, and the resulting responses were found to be accurate and consistent.
Schema Related Prompts
This section includes prompts focused on identifying, retrieving, and describing schemas within the platform. Content here will guide users in locating schemas by name, type, or associated connection, as well as discovering schema-level metadata such as owners, descriptions, tags, or classifications.
Use these prompts to cover queries related to schema discovery, schema search behavior, and schema-specific filtering or metadata lookup.
Get all details of a schema, for example, “Get me all details of xxxxx schema”.
Retrieve schema ownership details, for example, “Give me the owner of the Schema xxxxx”.
Identify schemas associated with a domain, for example, List all the Schemas associated with the Domain xxx.
Retrieve schema identification information, for example, “Give me the ID of the schema Project”.
View schema-related stewardship information, for example, “Who is the Steward for the Schema xxxx”.
Display schema popularity metrics, for example, “What is the Popularity score for the Schema xxx.xxxxx with Connection Name xxxx xxx”.
List all schemas present in the Data Catalog.
Retrieve schema information linked with specific projects, for example, “Give me details about the Project xxxx Term Curation”.
Table Related Prompts
This section includes prompts for searching, identifying, and reviewing tables across supported integrations. It allows users to retrieve table metadata, including type, schema, connection source, ownership, description, and governance context. These prompts support documentation validation, structural understanding, and enterprise-wide table inventory analysis.
Get all details of a table, for example, “Get me all details of the xxxxx table.
Retrieve a list of tables with specific custom fields, for example, “Give me the list of Tables that have Custom Field M and Table-xxx-1”.
Retrieve all tables present in the Data Catalog.
Retrieve all tables present in a specific schema and connection, for example, “List all the Tables present in the Schema xxxx.xxxx with Connection Name Retail Bank”.
Retrieve certification, caution, violation, and inactivity-related tables from connections such as sqlserver_finance.
Retrieve tables based on metadata scores and parameters, for example:
Find the importance and popularity score of the customer table.
Find the metadata curation score of the customer table.
Find the DQ score of the CUSTOMER table.
Find the row count of the CUSTOMER table.
Find the tables where certification status is Inactive from the sqlserver_xxxx connection.
Identify tables by association and characteristics, for example:
Find me the objects of type MATERIALIZED_VIEW.
Find the tables for which rating is more than 4.
Find the tables for which the reference count is 15.
Find the top 10 tables that have more number of columns.
Find the tables where user1 is the steward.
Find the tables of the demodb schema.
Find the tables and columns of the demodb schema.
Table Column Related Prompts
This section includes prompts for reviewing column-level metadata within tables. Users can search and extract column names, data types, constraints, and semantic attributes to understand table composition and field-level structure. These prompts assist in data modeling, schema inspection, and downstream system analysis.
Get all details of a table column, for example, Get me all details of the xxx_NUMBER table column.
Retrieve table names associated with specific columns, for example, “What is the Table Name of the Table Column Project”.
File Related Prompts
This section includes prompts for exploring file-based assets across data lakes and storage platforms. It supports the retrieval of file metadata, including file names, formats, storage paths, security context, and catalog registration details. These prompts assist in governing structured and semi-structured assets and enabling audit-ready file discovery.
Retrieve all details of files, folders, or buckets, for example:
Get me all details of the third-party_jars file (Folder).
Get me all details of the oe-public-dataset file (Bucket).
Retrieve file-related business and metadata details, for example:
What is the business description of the file xxx_15million.csv?
Find the metadata curation of the file xxx./.xxz.csv.
Retrieve quality and access permissions, for example:
Find the Access permissions of the file xx./.xxxz_15xxxn.csv.
Find the profiling status of the file xx./.xxxz_15xxxn.csv.
Retrieve file columns present within a file, for example, “Find the file columns available in the file xx./.xxxz_15xxxn.csv”.
Retrieve the ID of a file, for example, “What is the ID of the File data_xxx”.
File Column Related Prompts
This section includes prompts for accessing column-level details for files registered in the catalog. Users can review file schema elements, data types, and structural definitions to understand data stored within files. These prompts support validation of file ingestion, metadata enrichment, and lineage-ready file structure capture.
Retrieve files associated with a specific file column, for example, “Give me the File names which have this File Column xxxx”.
Report Related Prompts
This section includes prompts for discovering and analyzing reports and dashboards. It enables the retrieval of report metadata such as report names, business context, ownership, platform source, and publication details. These prompts support BI cataloging, governance validation, and enterprise reporting transparency.
Retrieve all details of a report, for example, Get me all details of the xxxxx Summary report.
Report Column Related Prompts
This section includes prompts to access and review fields and measures used within reports. It allows users to retrieve column names, business definitions, transformation logic (where available), and data source mapping. These prompts support semantic validation, BI lineage initiatives, and KPI interpretation accuracy.
Retrieve all details of a report column, for example, Get me all details of Tableau Online-customer_usecase_xxxx_test.xxxxe report column.
API Related Prompts
This section includes prompts for discovering API endpoints and related metadata. Users can extract API names, versions, purpose descriptions, platform source, and governance roles. These prompts aid in cataloging integration services and improving interface visibility across distributed architectures.
Retrieve all details of an API, for example, “Get me all details of xxxsonFunxxxHistoryV1.getxxxnxuxxxHistory API”.
API Attribute Related Prompts
This section includes prompts for retrieving API parameter and payload field details. It allows users to identify attribute names, formats, constraints, and metadata attributes associated with API structures. These prompts support interface analysis, request-response modeling, and integration consistency.
Retrieve all details of an API Attribute, for example, Get me all details of projects-sumxx-opexxx3.rexxxtionDatetime API Attribute.
Retrieve schemas based on table association, for example, “Find the schema name that has this table oetableext”.
Retrieve details related to data classification and workflow, for example:
Give me information about Data Classification Recommendations.
Explain the Service Desk Ticket Approval Workflow.
Retrieve project and data object associations, for example:
Find the associated tags of xxxxtable.
Find the associated tags of the xxxx_V table.
Find the associated terms of the xxxxxx_V table.
Term and Tag Related Prompts
This section includes prompts for exploring glossary terms and associated tags. It enables identification of business definitions, term relationships, usage context, and mapped catalog assets. These prompts support semantic consistency, data literacy initiatives, and taxonomy adoption across the enterprise.
Retrieve objects or items associated with a term, for example, List all the objects for which the Term Bolt is associated.
Retrieve terms and their metadata details:
List all the Terms in the application.
List all the Terms for which Forecasting is the Category.
List all the Terms for which STATUS is PUBLISHED.
Give the details of the Term Wire transfer.
Give the Business Description of the Term xxx transfer.
Give the STATUS of the Term xxx transfer.
List all the Terms for which User2 is the Steward.
Certification and Status-Based Search
This section includes prompts for filtering assets based on certification status, lifecycle stage, or validation category. Users can locate certified, draft, deprecated, or trusted objects to ensure compliance and governed asset consumption. These prompts ensure decision-making relies on verified and approved information sources.
Supported scenarios include
Retrieve certified objects
Retrieve cautioned objects
Retrieve violated objects
Retrieve inactive objects
Retrieve objects where the certification status is inactive
Examples
Find the certified tables from the sqlserver_xxxx connection
Find the tables that are cautioned from the sqlserver_xxxx connection
Find the tables that are violated from the Server_xxxx connection
Find the tables that are Inactive from the Server_xxxx connection
Find the tables where certification status is Inactive from the sqlserver_xxxx connection
Object Type and Structural Filters
This section includes prompts for restricting search results to specific object classes or structural types. It supports filtering across catalog objects such as schemas, tables, files, reports, APIs, or views. These prompts enable precise discovery and acceleration of enterprise asset navigation.
Examples
Find the top 10 tables which has more number of columns
Find the tables which has these columns associatxxxxxtids,associaxxxts,associatedtags
Governance Role-Based Search
This section includes prompts to retrieve assets based on stewardship, custodianship, and assigned owners. Users can identify responsible parties and understand accountability structures across the catalog. These prompts support governance tracking, operational clarity, and data accountability enforcement.
Examples
Find the tables where User1 is the steward
Find the tables where User1 is the owner
Find me the steward of table "xxxdataTable"
Who are the governance roles of the table "xxxdataTable"
Table Rating, Profile, and Quality Evaluation
This section includes prompts for retrieving data quality and profiling indicators associated with tables. It supports visibility into data distribution, profiling summaries, column statistics, and user ratings. These prompts assist in data reliability assessment, quality governance, and trust scoring.
Examples
Find the tables for which rating is more than 4
What is the rating of the table "xxxdataTable"
What is the data quality score of the table "xxxdataTable"
What is the Profile Status of the table "xxxdataTable"
What is the "Row count" of the table "xxxdataTable"
Find the row count of the CUSTOMER table
Documentation and Description Lookup
This section includes prompts for retrieving descriptive text, comments, and asset documentation. It supports accessing defined business context, usage notes, metadata annotations, and semantic guidance. These prompts enhance transparency, knowledge reuse, and catalog richness for end-users.
Examples
What are the descriptions of the table "xxxdataTable"
What is the source description of the table "customer_customer_demo"
Get me the columns of the table "pg_type"
Lineage, Reference Count, and Popularity
This section includes prompts to retrieve lineage-related metrics, reference counts, and popularity indicators. It enables visibility into object dependency levels, usage frequency, and downstream asset influence. These prompts support impact understanding, consumption analysis, and system complexity mapping.
Examples
Does table "xxxdataTable" have lineage
What is the Reference count of the table "xxxdataTable"
What is the Popularity of table "xxxdataTable"
What is the Importance of the table "xxxdataTable"
Origin and Authority Classification
This section includes prompts to identify the system of origin, authority ranking, and certified source context. It supports distinguishing trusted upstream datasets from secondary or derived sources. These prompts enhance traceability, reliability assessment, and authoritative data access.
Examples
Is the table UserLogins an Authoritative Source
Is the table UserLogins an Originative Source
Connection and Location Context
This section includes prompts for identifying the connection system, storage context, and environment location for catalog objects. It supports mapping assets to development, testing, or production systems and understanding platform integration lineage. These prompts strengthen structural clarity and operational oversight.
Examples
What is the connection type of "SQL Server xx Dec 2025"
Find the connection name of table xxxx
What is the database name of table xxxx
What is the schema name of table xxxx
Access Management and Object Activity
This section includes prompts for retrieving access controls, permission context, and usage activity. It supports understanding user access boundaries, viewing consumption patterns, and validating governance enforcement. These prompts contribute to security management and audit compliance.
Examples
Is access management enabled for table ApolloHospitals
What is the "Last Modified By" of table "xxxxdataTable"
Who created the table "xxxxdataTable"
Project Related Prompts
This section includes prompts for retrieving catalog projects, associated tables, tasks, and metadata initiatives. It supports viewing project titles, contributors, lineage alignment, and asset onboarding activities. These prompts enhance visibility into catalog adoption and structured metadata curation initiatives.
Retrieve ownership and association information, for example:
Who is the Owner of the Project My Access Cart?
List all the projects for which User1 is the owner.
Functional Behavior
Discovery Mode operates on metadata indexed from connected data sources and glossary repositories. When a prompt is executed, the system interprets the input, identifies relevant object types, and fetches metadata properties based on stored relationships.
Analytics Mode functions as an extension to Discovery Mode, enabling analysis on datasets brought into the workspace. It supports summarization, aggregation, and visualization-related prompts that are directly dependent on data retrieved from Discovery Mode.
Both modes are context-aware and function based on existing metadata coverage.
Prompts must be metadata-driven; analytical or conditional expressions outside the metadata scope are not processed.
The functional accuracy depends on the completeness of metadata available in the connected catalog sources.
Functional Notes
Discovery Mode Coverage
Supports metadata-level queries across Data Catalog objects, including Schemas, Tables, Columns, Files, File Columns, Reports, APIs, and Glossary terms.
Analytics Mode Coverage
Enables analytical insights and summarization based on data fetched through Discovery Mode.
Metadata Retrieval Scope
Focused on descriptive, ownership, association, scoring, and status-based metadata.
Prompt Format
Prompts must follow clear, object-oriented phrasing such as “Get me all details of [object]” or “Find [attribute] of [object].”
Validation
All prompts included in this document were validated successfully.
Test Data
Automated test validation was conducted using structured metadata from multiple data domains.
Limitations
Discovery Mode currently supports metadata retrieval and object-level exploration but does not support contextual, analytical, or multi-attribute metadata queries.
Functional Limitations
Discovery Mode supports only metadata-level retrieval and not analytical data operations.
Lineage, impact analysis, or object relationship visualization cannot be performed through prompts.
Queries requiring workflow data, approval stages, or custom UI interactions are not supported.
Query Interpretation Limitations
Prompts containing logical or comparative operators are not supported.
Multi-condition prompts combining metadata attributes (for example, steward and custodian) cannot be processed.
Custom field-based or formula-driven filtering is not recognized.
Metadata Scope Limitations
Derived or calculated metadata attributes cannot be generated dynamically.
Contextual dependencies between unrelated metadata entities are not identified.
Certain metadata types, such as external glossary relationships, may not be retrieved unless linked to a Data Catalog object.
Contextual Understanding Limitations
Prompts requiring synonym, context-based, or implicit meaning interpretation may return limited results.
Semantic associations that depend on business rules or natural language variations are not yet implemented.
Detailed Limitation Breakdown
Conditional Metadata Filtering
Discovery Mode cannot interpret or execute multi-condition queries involving multiple metadata attributes within a single prompt.
For example: Show tables where the steward is User1 and the custodian is User2.
Custom Field-Based Filtering
Queries involving metadata conditions or comparisons on custom fields are not supported. For example: Show tables where the custom field equals Click per Second.
Cross-Object Relationships
The system cannot establish relational links across different metadata objects within a single query. Queries that require cross-referencing objects, such as finding tables and reports simultaneously based on shared attributes, are unsupported.
Data-Level Queries
Discovery Mode is limited to metadata retrieval and does not fetch or analyze actual data values. Prompts that request data-level comparisons, aggregations, or statistical analysis cannot be processed.
Complex Analytical Context
Queries that require contextual understanding, such as business rules or domain-specific object relationships, are not supported.
Custom Attribute Derivation
Discovery Mode does not calculate derived values or perform runtime attribute derivation. Any metric requiring computation or aggregation based on existing metadata fields must be handled externally.
Non-Metadata Driven Searches
Discovery Mode cannot interpret prompts that search beyond the metadata scope, such as data lineage paths, impact analysis, or process-level traceability.
Advanced metadata filtering, contextual object relationships, and semantic search capabilities are expected to be available after integration with the planned RAG (Retrieval-Augmented Generation) architecture in future versions.
Bonus Prompt Category: Self-Service / Support
Useful for routing users to the Support Portal or help center.
Prompts:
How can I request access to a table?
Where can I raise a support request?
How to define a new glossary term?
Show me help articles on masking policies.
I want to learn more about lineage graphs.
Ownership Discovery Prompts
Who is the owner of the customer table?
Show me the steward for the term ‘Churn Rate’.
List all tables owned by John Smith.
Who governs the 'Sales Transactions' dataset?
Find glossary terms owned by the marketing team.
Which objects does Ramesh Sharma steward?
Which reports are managed by the finance department?
Copyright © 2025, OvalEdge LLC, Peachtree Corners, GA USA
Last updated
Was this helpful?