❄️ Mid-Winter 2026 Platform Update: Enhanced Joins, Connector Copilot Improvements, and HubSpot Certification

Hi there,

This update delivers significant enhancements to Nexadata's data transformation capabilities and AI-powered features. We've expanded join operations to support multi-stage workflows, improved Connector Copilot with better authentication handling and editing capabilities, and enhanced error transparency across AI features. We've also refined the platform experience with smarter UI behaviors and resolved several bugs affecting mapping groups, scheduled workflows, and connector reliability.

🔗 Enhanced Join Transformations with Supporting Pipeline Outputs

Nexadata now supports cross-pipeline joins, allowing you to reference materialized outputs from separate "supporting" pipelines during join operations. This unlocks complex multi-stage enrichment workflows that previously couldn't be handled within a single pipeline execution.

How It Works

When configuring a join transformation, you can now:

• Select outputs from one or more Supporting Pipelines as join sources and execution contexts
• Allows for building modular, reusable data enrichment workflows
• Handles edge cases requiring heavy pre-processing or lookup table generation

This capability enables scenarios such as multi-stage enrichments, complex lookups that require separate processing, and hierarchical data relationships spanning multiple execution contexts.

Use Cases

• Lookup Table Enrichment: Pre-process and materialize reference data in one pipeline, then join it efficiently across multiple downstream pipelines
• Multi-Stage Transformations: Break complex transformations into modular steps, each with its own pipeline, then combine results through cross-pipeline joins
• Hierarchical Data Processing: Build parent-child relationships across pipeline boundaries

🧠 Connector Copilot: Multiple Authentication Methods

The Connector Copilot now automatically detects and supports APIs with multiple authentication schemes, significantly expanding the range of systems you can connect to using natural language.

Previously, Connector Copilot could only handle APIs with a single authentication method. Now, when analyzing an OpenAPI specification, it detects all available security schemes and prompts you to select the appropriate method for your use case.

Supported Scenarios

• Multiple API Keys: Systems like Stripe or Pigment that offer different key types (standard API keys, restricted keys, or organization-level keys) for different access patterns
• Mixed Authentication Methods: APIs that support both API key and OAuth 2.0 authentication, allowing you to choose the method that best fits your integration needs
• Flexible Security Configurations: Choose the authentication method that best fits your security requirements

When configuring a connection, Connector Copilot will automatically present available authentication options and guide you through the selection process.

✏️ Connector Copilot: Edit Datasets, Connections, and Systems

You can now edit datasets, connections, and external systems created through the Connector Copilot and align them with manually configured resources across the platform.

Previously, Connector Copilot-created resources were immutable after creation, requiring users to delete and recreate them for any modifications. This update introduces:

• Dataset Editing: Modify data source configurations, update field mappings, and adjust pagination settings
• Connection Editing: Update authentication credentials, change API endpoints, and reconfigure connection parameters
• External System Management: Allows users to add a logo to a system created via Connector Copilot, providing a quick and easy visual reference when building Pipelines

This change eliminates the need to recreate resources from scratch when requirements change, streamlining iteration and configuration management.

🔍 Enhanced AI Error Feedback and Transparency

AI Copilot now provides detailed logs and actionable error messages when transformations fail, significantly improving debugging and iteration speed.

Previously, AI Copilot failures resulted in generic error messages with limited diagnostic information. Users had no visibility into why a transformation failed or how to adjust their prompts for better results.

What's New

• Advanced Error Logs: View detailed error messages that explain what went wrong during AI-assisted transformation generation
• Actionable Feedback: Receive specific guidance on how to adjust prompts or configuration to resolve issues
• Debugging Context: Access raw error details to quickly diagnose and iterate on complex transformations

Additionally, we've softened the UI response to transient AI failures. Instead of displaying harsh red error indicators, the platform now suggests retry options with gentler messaging, recognizing that many AI failures are temporary and resolve on retry.

🎨 Platform Improvements

Column Rename Propagation

Column renames now automatically propagate to all downstream transformations that reference them. When you rename a column in one transformation, every subsequent transformation in the pipeline updates automatically, maintaining pipeline integrity without manual intervention.

Unified Ellipsis Menus

Mapping Groups and Pipelines now feature consistent ellipsis menu options (Edit, Delete) across the platform, matching the behavior of Datasets and Connections. This creates a consistent, predictable interface experience across Nexadata.

Improved Search Field Behavior

Property and column selection fields now automatically clear search text after selection, eliminating the need to manually delete text before performing another search. This small refinement significantly improves workflow efficiency when configuring transformations.

🧱 Platform Fixes

• ENG-845 – Bug Fix: Connector Copilot now provides clear, actionable feedback when API pagination limits are exceeded, replacing generic error messages with specific validation guidance.
• ENG-842 – Bug Fix: Resolved issue where the Dataset Builder workflow would fail after selecting an API endpoint during Connector Copilot dataset creation.
• ENG-840 – Bug Fix: Scheduled workflows marked as "active" now fire as expected; resolved execution reliability issue affecting workflow automation.
• ENG-734 – Bug Fix: Property selection fields now automatically clear search text after selection, streamlining the configuration workflow.
• ENG-827 – Bug Fix: Save button in the Mapping Group editor remains visible when scrolling through long rule lists, preventing accidental loss of configuration work.
• ENG-831 – Bug Fix: Mapping group rules now correctly handle exact value matching without type conversion issues (e.g., "400000" no longer appears as "400000.0").
• ENG-826 – Bug Fix: Conditional mapping groups now correctly evaluate rules when referenced columns contain blank values.

🎉 Nexadata Certified on HubSpot Marketplace

We're excited to announce that Nexadata is now certified on the HubSpot Marketplace, making it easier for HubSpot users to discover and integrate Nexadata into their data workflows.

This certification validates Nexadata's integration quality and security standards and reflects our commitment to providing seamless connectivity with leading business platforms. As part of this certification process, we've refined our HubSpot OAuth implementation to align with HubSpot's best practices for scope minimization and data access.

📄 View Nexadata on the HubSpot Marketplace →

🔭 What's Next

We're continuing to enhance Connector Copilot and expanding natural language capabilities across the platform. Here's what we're working on:

• Multi-Request API Support in Connector Copilot: Automatically handle APIs that require multiple sequential requests to retrieve complete datasets, including async job-based extractions where data preparation happens server-side
• Metadata-Aware Dataset Configuration: Connector Copilot will automatically discover and use metadata APIs to populate dropdown options and provide guided field selection during dataset creation, reducing manual configuration
• Natural Language Mapping Groups: Create and iteratively edit standalone mapping groups entirely through natural language prompts, bringing full parity with pipeline natural language workflows
• Holistic Pipeline Editing with Artifacts: Refine and revise entire pipelines conversationally with artifact support, enabling broader iterative changes beyond transformation-by-transformation updates
• Pipeline Versioning: Automatic snapshot creation upon execution to track how pipelines looked at the time of each run, enabling better audit tracking and troubleshooting

Thanks for being part of the Nexadata journey.

— The Nexadata Team