September 2025 Summaries

Migrating SAP data to Snowflake can be difficult because SAP environments are complex, business-critical, and often supported by inefficient legacy ETL processes, while CData Sync is presented as a no-code platform for secure near-real-time replication. The recommended approach begins with defining analytical goals, prioritizing valuable data domains, setting measurable success criteria, and documenting regulatory requirements. Organizations should then prepare SAP connectivity through ODP or RFC endpoints, dedicated least-privilege service accounts, and controlled network access. The guidance favors a hybrid strategy that combines an initial bulk historical load with ongoing incremental replication, balancing migration speed with current data availability. CData Sync configuration includes secure authentication, scoped table selection, schema-change detection, encryption, retries, alerts, and a transition from batch loading to incremental updates. After deployment, teams should validate row counts and checksums, test latency, monitor pipeline and warehouse health, establish threshold-based alerts, and scale Snowflake or replication instances as demand grows, while using audit logs and access controls to support GDPR and SOC 2 compliance.

Integrating Microsoft SharePoint with Snowflake can make documents, metadata, and permissions available for analytics, with CData Sync presented as a no-code platform for automated replication from SharePoint and more than 100 other source types. The process begins by defining file coverage, metadata, versioning, access controls, and compliance needs, then securely configuring the SharePoint connector through OAuth 2.0 or Kerberos and preparing a dedicated Snowflake database, schema, role, and warehouse. Data mappings can align SharePoint fields such as file names, modification dates, and permissions with Snowflake columns, while incremental synchronization uses timestamps to transfer only changed records and reduce API usage. Ongoing monitoring through CData Sync dashboards and Snowflake query history helps validate loads, performance, and row counts, while permission data stored as JSON can support Snowflake row-access policies. The guidance also addresses common problems including authentication failures, API throttling, schema drift, and stalled jobs, and recommends scaling with multi-cluster warehouses or bulk loading as data volumes grow.

CData Connect AI is a cloud-based platform designed to give AI assistants such as ChatGPT, Claude, and Microsoft Copilot secure, real-time access to proprietary enterprise data from more than 270 sources, including CRM, ERP, finance, support, databases, and SaaS applications. Using CData’s Model Context Protocol, it acts as a universal adapter that translates natural-language questions into live data queries without requiring code, pipelines, local installation, or data duplication. The platform emphasizes current, governed results to reduce reliance on outdated information and potential AI hallucinations, while offering role-based controls, source-permission enforcement, audit trails, OAuth 2.1, AES-256 encryption, SOC 2 certification, and GDPR compliance. It supports departmental use cases such as analyzing sales delays, support trends, campaign ROI, and financial performance, and can be deployed from individual analysts to large organizations through a no-code setup process and a managed cloud service with a 14-day free trial.

CData announced its first virtual CData Connect AI Hackathon, a global competition beginning September 25, 2025, that invites developers, analysts, and innovators to create AI agents, copilots, or assistants powered by live enterprise data. Participants must use CData Connect AI, a no-code platform that provides governed access to hundreds of systems including Salesforce, HubSpot, QuickBooks, and Snowflake, alongside an MCP-enabled AI tool to demonstrate a practical use case. Entries are due October 24, 2025, and should describe the project, connected systems, AI integrations, and business problem addressed, with optional screenshots or demonstrations. A CData expert panel will assess submissions for creativity, usefulness, functionality, and presentation, awarding gift cards of $500, $250, and $100 to the top three projects. Suggested ideas include customer-support assistants, automated Snowflake insight sharing, finance-reporting copilots, and marketing recommendation tools using real-time business data.

CData Connect AI is presented as a solution for reducing AI hallucinations in enterprise settings by giving large language models such as ChatGPT and Claude governed, real-time access to authoritative business data from systems including Salesforce, NetSuite, HubSpot, Google Ads, and Sage Intacct. The approach aims to prevent inaccurate responses caused by models lacking current organizational context, while allowing IT teams to control data access, monitor queries, support compliance, and limit unsanctioned AI use. Unlike traditional ETL and data warehouse pipelines, which can introduce latency and require data duplication, Connect AI provides direct query-based access to source systems, enabling more current, context-aware responses with lower infrastructure overhead. The platform is also described as supporting no-code exploration for business users while preserving IT governance, with the goal of building trust in AI-generated insights and enabling safer enterprise adoption.

CData has renamed Connect Cloud to CData Connect AI, positioning the existing data virtualization platform as a managed Model Context Protocol platform for connecting enterprise data securely and in real time to AI assistants, agents, and workflows. Existing customers retain their connectors, credentials, configurations, pricing, and core capabilities, while gaining MCP support, a Universal MCP endpoint, and native integrations with tools such as ChatGPT, Claude, Microsoft Copilot Studio, CrewAI, LangChain, and other compatible models. The platform is designed to let AI query live SaaS, database, ERP, and other business data without ETL, replication, or custom code, while maintaining source-level permissions, encryption, auditing, and row-, column-, and table-level access controls. Suggested uses include live sales and marketing analysis, financial forecasting, secure IT automation, and AI-powered product features, with support for private or self-hosted models that can use MCP or compatible APIs.

CData Sync’s Q3 release expands support for modern lakehouse, legacy database, and operational data workflows through open Delta Lake tables, enhanced change data capture, and additional Reverse ETL destinations. Native Delta support across Amazon S3, Azure Blob Storage, Google Cloud Storage, Azure Data Lake Storage, and Open Mirroring for Microsoft Fabric OneLake provides transactional writes, schema management, and compatibility with tools including Databricks, Spark, Trino, Fabric, and Power BI. The release adds CDC support for IBM DB2 AS400/iSeries and improves MySQL replication, while hard-delete tracking across CDC sources helps downstream platforms maintain an accurate source-of-record state. Reverse ETL now includes Sage Intacct, Veeva Vault CRM, Pardot, and Salesforce Marketing Cloud, allowing finance, life sciences, and marketing teams to send warehouse data to operational systems. A new Salesforce_Formulas schema also replicates only calculated and roll-up fields with primary keys, reducing unnecessary Salesforce data while preserving business logic for analytics.

CData has launched Connect AI, a managed Model Context Protocol platform designed to give AI assistants, agents, workflows, and embedded applications governed, real-time access to more than 300 enterprise data sources. The company argues that many enterprise AI initiatives fail because AI lacks secure access to the structured, interconnected business data held across systems such as Salesforce, Jira, and NetSuite, while existing approaches such as third-party MCP servers, RAG implementations, and custom integrations can introduce security, development, or maintenance challenges. Built on CData’s connectivity infrastructure, Connect AI accesses data in place rather than replicating it, preserves metadata and relationships between business objects, supports both data queries and actions through a unified connection, and inherits source-system role-based permissions with options for additional controls. The platform can be cloud deployed or embedded by independent software vendors, allowing them to offer white-label, self-service connections between customer data and AI features without building and maintaining their own connectors.

CData Arc v25.3, released in Q3 2025, adds AI-assisted XML and EDI mapping powered by OpenAI or locally hosted Ollama models, designed to suggest transformation logic and reduce the time and expertise required for partner onboarding and custom integrations. The release also introduces an ICAP Connector for integrating antivirus scanning, content filtering, and data-loss prevention into managed file transfer workflows, with TLS, configurable HTTP encapsulation, and authentication support. Security updates include a unified SSO configuration experience for Microsoft Entra ID and Okta, simplified FIPS-compliant deployment, and HMAC signature authentication for incoming webhooks in Professional and Enterprise editions. Usability improvements include a redesigned XML Map interface, a reorganized Settings dashboard with centralized security, licensing, and usage controls, and message views that emphasize current header values while retaining historical records for troubleshooting and audits.

Database replication copies and synchronizes data from a source to one or more targets, supporting continuous or scheduled updates for high availability, disaster recovery, analytics offloading, multi-cloud sharing, AI/ML workloads, and cloud migrations. Unlike backups and batch ETL, replication commonly uses change data capture (CDC) to deliver inserts, updates, and deletes within seconds or minutes, with log-based CDC generally offering the lowest source-system impact. The guide compares synchronous replication, which prioritizes consistency and zero data loss at the cost of latency, with asynchronous replication, which improves responsiveness but can temporarily lag, as well as snapshot, merge, and transactional approaches. It recommends evaluating tools by connector breadth, CDC capabilities, performance features, pricing structure, security controls, governance, schema evolution, monitoring, and conflict-resolution support for bi-directional environments. It presents CData Sync as a no-code option for heterogeneous on-premises and cloud systems, emphasizing its broad connector catalog, log-based CDC support, connection-based pricing, encryption, role-based access, auditing, performance tuning, automated recovery, and integrations for monitoring and scalable replication workloads.

Connecting Jira to Google Sheets can replace manual CSV exports and fragile scripts with live or scheduled access to issue data, using JQL to retrieve targeted records for reporting and analysis. The guide compares five approaches: CData Connect AI for live, governed no-code connectivity; Atlassian’s official Jira Cloud for Google Sheets add-on for simpler scheduled imports; two-way synchronization tools configured for one-way reporting; automation platforms such as Zapier, Make, and n8n for event-driven workflows; and Google Apps Script with the Jira REST API for maximum customization. It presents CData Connect AI as best suited to teams requiring real-time data, centralized security controls, SSO or OAuth, granular permissions, audit logging, and low maintenance, while noting that other options may fit smaller datasets, periodic reporting, multi-system workflows, or developer-specific requirements. Selection should depend on required data freshness, scale, hierarchy and custom-field handling, governance needs, reliability expectations, maintenance capacity, and total cost of ownership.

CData Foundations AI Day on Sept. 24 will convene speakers from organizations including Google, ServiceNow, AWS, Ellie.ai, Sisense, and Ataccama to discuss why enterprise AI success depends on infrastructure, interoperability, governance, and trustworthy data rather than models alone. Sessions will examine multi-agent collaboration across platforms, live metadata and AI-native data modeling, embedded analytics with real-time connectors, and ways to deliver AI value before full data centralization is complete. Speakers will also address safe and compliant autonomous agents, lessons from large-scale generative AI deployments, and data-quality practices that unify fragmented sources and support governed access. The event presents connected data, models, systems, and workflows as essential for moving enterprise AI initiatives beyond isolated pilots toward reliable operational results.

CData’s ninth Vibe Querying episode features FP&A Director Elliot York demonstrating how the company’s Salesforce MCP Server helps revenue operations teams analyze complete customer journeys through conversational AI. Model Context Protocol connects AI clients such as Claude and Gemini to live business data across CData’s library of more than 350 sources, enabling users to ask natural-language questions without detailed schema knowledge or prebuilt reporting pipelines. York developed reusable “purposeful prompts” to examine closed-won deals, stakeholder roles, and interactions preceding opportunity creation, allowing Claude to combine Salesforce opportunities, contacts, emails, activities, support cases, and custom objects into chronological deal timelines and qualitative insights. The approach identified factors such as partner influence, stakeholder dynamics, product-fit evolution, and historical account relationships that can be difficult to find through manual CRM research. According to the episode, this reduced customer-journey analysis from hours of Salesforce navigation to roughly 15–30 minutes, and CData has begun incorporating the findings into weekly go-to-market discussions. The server’s visibility into underlying SQL queries also allows finance and operations users to validate data access and analysis methods, while York sees future potential for scaling the process from individual deals to cohort-level analysis across thousands of customers.

CData Foundations 2025’s Data & Analytics Day, scheduled for September 17, will examine how organizations can address stale dashboards, delayed reporting, fragmented data systems, and governance bottlenecks that undermine confidence in analytics and limit AI adoption. The event will feature customer examples from NCCO, which is transitioning legacy ERP data to Snowflake for digital products, and Red Wing Shoes, which redesigned data flows to support point-of-sale reporting across hundreds of stores. Sessions from CData, Argano, BearingPoint, Databricks, and FinThrive will discuss modernization strategies including semantic layers, data ingestion, governance, managed platforms, and architecture choices that improve data access and reliability. The program emphasizes that strong analytics foundations are necessary both for dependable business reporting and for AI initiatives, especially as unreliable or inaccessible data contributes to poor outcomes from generative AI pilots.

CData Connect Spreadsheets is presented as a no-code tool for connecting Salesforce data to Google Sheets in real time, aiming to replace manual CSV exports and reduce reporting delays, version inconsistencies, and errors. The process involves installing the Sheets extension, creating and authorizing a Salesforce connection through OAuth, selecting a production or sandbox environment, and building queries through a visual interface or SQL-based options. The guidance emphasizes filtering data at the Salesforce source, limiting selected fields, using pagination and row limits, and scheduling refreshes carefully to maintain performance and stay within Salesforce API quotas. Connected data can support Sheets features such as pivot tables, charts, formulas, automated refreshes, alerts, and blending with sources such as analytics, advertising, or billing platforms. For write-back use cases, it recommends starting with read-only access and applying role-based controls, sandbox testing, field mappings, audit logs, and rollback procedures to govern updates to Salesforce.

CData’s Vibe Querying episode featuring product manager Jonathan Hikita demonstrates how its Salesforce MCP Server uses Anthropic’s Model Context Protocol to let AI clients such as Claude access, analyze, and update Salesforce data through natural-language prompts. Faced with scattered use-case information across Salesforce opportunities, emails, activity notes, and support cases, Hikita used the system to identify new-business opportunities lacking documented use cases, extract structured details about data sources, connecting technologies, destinations, and business purposes, and write the results back into Salesforce custom objects. The approach reduced a process that previously required weeks of manual review to a few hours, while preserving knowledge for sales, marketing, and product teams. Analysis of the resulting use cases also surfaced strategic patterns, including demand for QuickBooks-to-Power BI integrations amid Microsoft’s planned deprecation of its native QuickBooks connector, as well as HubSpot analytics and SAP HANA-to-Databricks opportunities. The example highlights CData’s ability to work with Salesforce custom objects, validate data formats, navigate relationships, and support both read and write operations, positioning conversational AI as a way for product teams to focus more on market decisions and less on manual data compilation.