May 2025 Summaries

Data integration can involve either physically copying data through ETL, ELT, or change-data-capture pipelines or virtually connecting to source systems in real time. CData Connect AI is presented as a platform for creating secure virtual datasets that allow tools such as spreadsheets, BI platforms, and automation applications to query live source data without maintaining replicas, warehouses, or synchronization jobs. This approach can reduce costs, accelerate setup, preserve data freshness, and limit governance risks by keeping sensitive information within existing systems and compliance boundaries. Live connections are positioned as particularly suitable for ad-hoc reporting, self-service analytics, and automations that require current records, while data replication remains useful for large historical analyses, machine learning, disaster recovery, and intensive transformations. The recommended strategy is often a hybrid architecture, using Connect AI for immediate access to operational data and CData Sync to archive, transform, or replicate data for long-term analytical needs.

CData Connect AI is presented as a SaaS connectivity platform designed to help product owners and project managers use live Jira data for resource planning, cost analysis, KPI reporting, and capacity forecasting without relying on fragile exports or custom scripts. Through no-code OAuth setup, users can query, join, sync, and in some cases update Jira data from tools including Excel, Google Sheets, Power BI, Tableau, Looker, and automation platforms, while also combining Jira with sources such as Tempo, GitHub, finance systems, and support data. The platform supports use cases such as forecasting team capacity, calculating cost per story point from logged hours and payroll information, and creating automatically refreshed dashboards for velocity, cycle time, lead time, and defects. Users connect Jira, select needed data, and access it through SQL, OData, ODBC, or JDBC endpoints without installing drivers or maintaining servers, with a free trial offered for initial setup.

CData positions its Databricks Integration Accelerator as a no-code alternative to code-intensive native ingestion tools such as Delta Live Tables, Lakeflow Connect, and Azure Data Factory, which can require custom API integrations, Spark transformations, schema-evolution handling, and change data capture logic. The toolkit supports replication from more than 270 enterprise systems into Delta Lake, automated metadata and datatype management, built-in CDC and schema evolution, live federation through Lakehouse Federation, and governance through Unity Catalog. The text argues that broader business-system connectivity can reduce delays, maintenance burdens, data-quality risks, and governance gaps that affect analytics initiatives such as Customer360, marketing analysis, and financial reporting. NJM Insurance is presented as a customer example, using CData Sync to consolidate data from more than 10 advertising platforms into its Databricks lakehouse, reportedly reducing development time from 20 days to two days per source and achieving 66% cost savings while expanding support for risk modeling and fraud detection.

CData’s “Vibe Querying with MCP” episode demonstrates how product managers can use natural-language conversations with Claude and CData’s MCP Server for Salesforce to analyze support cases and prioritize product-roadmap work. In the episode, CData product management director Jaclyn Wands examines more than 1,400 Sync support cases, moving from a broad analysis of recurring authentication and connection problems to specific configuration issues, particularly SQL Server Windows authentication. The process produces development-ready recommendations, including example cases, technical fixes, estimated implementation effort, and expected reductions in support tickets. Opportunities are ranked by effort and impact, with suggested quick wins such as default timeout configuration and automated trial-license extensions, alongside larger initiatives involving license and communication management. The presentation emphasizes “purposeful prompting,” source validation, and reusable data-table maps to improve accuracy and speed for future analyses, positioning MCP-enabled conversational querying as an alternative to manual reporting, scripting, and data cleanup that can support cross-functional decision-making and measurable ROI.

Beginning in April 2026, HMRC’s Making Tax Digital initiative will require VAT-registered businesses and most self-employed professionals to maintain digital records and submit returns electronically, creating potential disruption for organizations reliant on spreadsheets or paper records. CData positions its Connect Spreadsheets product as a way to meet these requirements while continuing to work in Microsoft Excel or Google Sheets, using connectors to accounting platforms including QuickBooks, Sage Intacct, Xero, and NetSuite. The service supports importing live accounting data into spreadsheets, uploading updates back to source systems through two-way synchronization, and applying security controls such as encryption and granular permissions. It also promotes quick setup without coding or consultants, the ability to combine accounting information with data from sources such as Shopify and inventory systems, and future integration with analytics and automation tools. Users can begin with a free account and add-in, authenticate an accounting source, import or update records, refresh data as needed, and maintain digital VAT records for filing.

Model Context Protocol (MCP), a standard defined by Anthropic, is presented as a way to give AI models secure, real-time access to business systems, data, and tools that conventional prompts alone cannot provide. Large language models can produce inaccurate or fabricated responses when they lack current details such as inventory changes, customer interactions, or analytics settings, forcing users to manually supply incomplete and potentially outdated context. By connecting models directly to trusted business data sources, MCP can support activities such as reporting, inventory updates, sales management, and scheduling while replacing one-off integrations and manual context sharing with a standardized, scalable approach. The approach aims to move AI from isolated, generic outputs toward more relevant, reliable decision support based on live organizational information, with CData MCP Servers cited as one implementation option.

Retrieval-Augmented Generation (RAG) and the Model Context Protocol (MCP) are complementary technologies for connecting large language models to external information and systems. Introduced by Facebook AI Research in 2020, RAG retrieves relevant content from sources such as documents, knowledge bases, and vector databases to ground model responses, reduce hallucinations, and incorporate current or domain-specific knowledge. Released in November 2024, MCP standardizes how models access structured and unstructured sources, including APIs, databases, and resources, while also enabling controlled actions through a client-server architecture and access-management features. RAG is primarily suited to semantic retrieval of bulk unstructured context for generation, whereas MCP supports live data exploration, structured context delivery, and interactions with connected systems. Used together, they can support applications such as customer support, sales intelligence, and risk advisory agents by combining retrieved text-based knowledge with real-time customer, operational, or compliance data.

CData Sync is presented as a flexible alternative to Fivetran that can run alongside existing Fivetran pipelines, allowing teams to test or migrate workloads gradually rather than undertake a full replacement. It supports replication to the same warehouse destinations, offers control over scheduling and schemas, and may be used permanently with Fivetran retained for selected edge cases or redundancy. The platform emphasizes hybrid and on-premises deployment, near-real-time change data capture without requiring database owner privileges, and consolidation of custom scripts or legacy jobs. CData Sync also contrasts its connection-based fixed annual pricing with Fivetran’s usage-based model, positioning itself as a potentially more predictable option for high-volume data movement.

Informatica’s planned end of software maintenance for PowerCenter is prompting organizations to reconsider legacy ETL systems, while the passage argues that cloud-only replacements such as IDMC, Fivetran, and Matillion can create challenges involving infrastructure control, regulatory compliance, data exposure, and usage-based costs. It presents CData Sync as a hybrid data replication alternative that can operate on premises, in private clouds or VPCs, or as managed private-cloud SaaS, enabling teams to retain control over sensitive data and adopt cloud services gradually. The platform is described as supporting firewall-protected and air-gapped deployments, internal security standards, batch and incremental replication, change data capture, SQL transformations, dbt compatibility, scheduling, and observability. Its connection-based flat pricing is positioned as more predictable for high-volume workloads, and examples involving GSK, NJM Insurance, Manhattan Associates, and Recordati are cited to illustrate use in regulated, hybrid, and self-managed environments.

CData Sync is presented as an alternative to Fivetran for organizations seeking more deployment flexibility, predictable connection-based pricing, enterprise security, broad source and destination support, and built-in or dbt-based transformations. The migration process begins with auditing and exporting Fivetran connector settings, schemas, transformations, schedules, schema-drift behavior, and downstream dependencies, ideally documented in a mapping blueprint. Teams then install or access Sync, configure source and destination connections, rebuild replication jobs with selected tables, optional transformations, CDC, and matching schedules, and test results through logs, previews, record counts, and production reporting. After validating pipelines, organizations can schedule a low-usage cutover, redirect downstream tools, disable Fivetran connectors, and use Sync monitoring and workspace features to optimize operations, separate environments, and replicate configurations across teams or clients.

CData has expanded its ADP integrations across Connect AI, Connect Spreadsheets, and Sync, enabling HR teams to access live payroll, workforce, benefits, and timekeeping data in tools such as Excel, Google Sheets, Power BI, Tableau, data warehouses, Power Automate, and Zapier. The integrations are positioned as an alternative to manual exports, spreadsheet cleanup, and IT-dependent data uploads, with automatic refreshes intended to support faster reporting, real-time monitoring of payroll accuracy, overtime, and attrition, and automated workflows for tasks such as discrepancy alerts and cost-center reporting. Setup involves obtaining ADP API Central access, authenticating through OAuth, selecting ADP data, and choosing a destination tool. CData also cites enterprise security features including role-based access, SOC 2 compliance, and encryption, and highlights Advantage Investigations’ reported use of ADP, CData, and Power BI to accelerate payroll approvals and reduce timecard fraud.

Fivetran’s acquisition of reverse ETL provider Census is presented as evidence that activating warehouse data in business applications has become a central part of modern data integration, enabling teams to send enriched data to tools such as Salesforce, HubSpot, and Dynamics 365. The source notes potential concerns that Census could eventually adopt Fivetran’s usage-based pricing approach, potentially increasing costs for high-frequency or high-volume reverse ETL workloads, based on Fivetran’s earlier integration of HVR. It positions CData Sync as an alternative platform that combines ETL/ELT and reverse ETL, supports more than 350 data sources, provides transformations and dbt integration, and can be deployed on premises, in customer clouds, or through CData’s managed private cloud. CData Sync is described as using connection-based rather than volume-based pricing and as offering unified scheduling, transformations, logging, and pipeline management for data movement in both directions.

CData has introduced MCP Servers, an implementation of Anthropic’s Model Context Protocol designed to connect compatible large language models to enterprise systems such as Salesforce, Jira, and NetSuite through standardized SQL-based interfaces. The servers allow AI tools to discover data structures, query information across multiple sources, and perform supported write operations or stored procedures, potentially reducing reliance on static dashboards, custom integrations, and business-intelligence requests. The initial beta supports 25 enterprise data sources and Claude Desktop on Windows, with plans to add hundreds of connectors, Mac and additional LLM compatibility, cloud-hosted deployment, and other agent frameworks. CData says its existing connector technology provides performance features including query pushdown and bulk operations, while retaining data in source systems and honoring existing user authentication and permissions; it also cites SOC 2 Type II, ISO/IEC 27001:2022, and GDPR compliance.