June 2025 Summaries

CData reports rapid early adoption of its beta Model Context Protocol servers in the six weeks after launch, recording 294 downloads by 255 users across 139 companies and substantial growth toward the end of the period. Popular connectors included Google Sheets, NetSuite, QuickBooks, SQL Server, Salesforce, Excel, SAP, Google Analytics, and Power BI, reflecting demand across spreadsheets, enterprise applications, databases, and analytics systems. Telemetry indicated more than 26,000 queries, 2 million processed rows, 188 deployment nodes, and 57 actively used data sources, with Salesforce, SQL Server, Google Sheets, BigQuery, and Azure DevOps showing notable activity. The predominantly read-heavy workload, with roughly 20 reads for every write, is presented as evidence that users are employing AI assistants primarily for real-time data exploration, analysis, and insight generation rather than automation alone. CData argues that its SQL-capable, in-place connectivity provides broader analytical access than API-limited alternatives, and expects expanding production deployments, more varied sources, and new AI-native data architectures as its library of more than 350 connectors gains adoption.

Salesforce Connect provides real-time access to external data in Salesforce, but its per-source licensing, starting at $48,000 annually, can create substantial costs and integration complexity for organizations connecting multiple systems. CData API Server is presented as a unified OData gateway that aggregates databases, warehouses, and SaaS platforms into one API endpoint recognized by Salesforce Connect as a single external source. This approach could allow a company connecting four sources to reduce annual Salesforce Connect licensing from an estimated $192,000 to $48,000 while retaining real-time access to the same data. The platform also centralizes access management, allows additional sources to be added without Salesforce reconfiguration, and reduces the need for separate adapters or custom integrations. CData positions its API Server as an alternative to internally built integration layers, citing enterprise security, support for more than 100 data sources, and optimized real-time performance.

Sage Intacct API limits can constrain growing organizations that connect accounting data to dashboards, reporting tools, reconciliation workflows, and other systems, potentially leading to costly upgrades for additional capacity. CData Connect AI addresses this issue by caching Sage Intacct API results in an organization-controlled PostgreSQL database, scheduling refreshes according to business needs, and using incremental loading to retrieve only changed data. Managed through its Jobs feature without custom scripting, the approach allows tools such as Power BI, Tableau, and Excel to access current cached data while reducing direct API traffic. AMI Expeditionary Healthcare reportedly used this configuration to maintain frequently refreshed Power BI dashboards, avoid API rate limits and higher usage tiers, and reduce maintenance complexity.

Corporate FP&A teams often struggle to consolidate siloed data from ERP, accounting, and trading systems into Snowflake because legacy integration tools require manual mapping, create reporting delays, scale poorly, and pose compliance concerns. CData’s Snowflake Integration Accelerator aims to automate these workflows through no-code ingestion, change data capture, incremental replication, and enterprise security, with customer examples citing shorter job times and greater pipeline automation. Its three toolkits support faster ingestion from more than 270 systems, live access to Snowflake data through applications such as Power BI and Microsoft reporting tools, and structured and unstructured data integration for Snowflake Cortex AI use cases including forecasting and anomaly detection. The offering also emphasizes hybrid deployment flexibility, broad connectivity across enterprise applications, native Snowflake ecosystem support, and predictable licensing to help financial organizations improve the speed, reliability, and governance of analytics workflows.

Hybrid cloud environments can limit organizations’ ability to use Snowflake effectively because sensitive and mission-critical data often remains in isolated on-premises systems such as SAP, Oracle, and custom applications, creating challenges around data silos, compliance, latency, and brittle manual pipelines. CData positions its Snowflake Integration Accelerator as a no-code, enterprise-focused solution for connecting these systems to Snowflake, citing customer examples in manufacturing and healthcare where CData Sync automated ingestion, reduced manual maintenance, shortened job runtimes, and improved data availability. The accelerator includes ingestion capabilities supporting change data capture, Iceberg Tables, and more than 270 sources; live connectivity for applications and Microsoft tools; and integrations intended to support Snowflake Cortex AI workflows. CData also emphasizes deployment flexibility, direct integrations across cloud and on-premises ecosystems, predictable pricing, and compatibility with analytics, ETL, and reporting platforms as benefits for enterprises seeking faster hybrid data integration and analytics.

Gartner Data & Analytics Summit sessions emphasized that data leaders must balance faster delivery, decentralized ownership, and strong governance as complexity, talent shortages, and AI-driven demands increase. Analysts advocated federated operating models built around domain-aligned fusion teams, supported by shared platforms and enterprise governance, rather than relying on centralized data management teams to scale. They also identified an execution gap in which organizations may align data-product strategies with business goals but lack reusable capabilities, clear ownership, documentation, discoverability, and service-level expectations needed for broad adoption. Combining data mesh ownership practices with data fabric automation and metadata management was presented as one approach to scaling delivery. Master data management was portrayed as a continuing strategic foundation for defining, governing, and improving trusted data, with AI and automation helping accelerate data matching, enrichment, and quality while preserving oversight across distributed teams.

Snowflake’s May 2024 general availability release of Cortex AI provides native machine learning and large language model capabilities, but its effectiveness depends on organizations integrating diverse, high-quality data into Snowflake. The text argues that many enterprises leave substantial data unused because of fragmented pipelines, unstructured-data complexity, inadequate metadata management, scalability constraints, and manual processes. CData’s Cortex AI integration toolkit is presented as a solution that ingests structured and unstructured sources, automates metadata discovery, supports real-time ingestion and change data capture, and connects data directly to Cortex AI for uses including predictive analytics, automated documentation, and sentiment analysis. Examples from healthcare and manufacturing describe reduced ingestion times, pipeline automation, real-time insights, and inventory-related cost savings, while CData emphasizes no-code and low-code tools intended to reduce implementation effort for AI-ready Snowflake data pipelines.

CData Sync’s Q2 2025 release adds a SuiteAnalytics source connection for NetSuite, change data capture support for SAP ERP, and stronger authentication controls through two-factor authentication and SAML-based single sign-on. The NetSuite connector accesses the platform’s reporting engine to replicate saved searches, calculated fields, and joined datasets, complementing the existing API connection for raw transactional data and enabling teams to choose between detailed auditing data and reporting-ready datasets. SAP ERP CDC provides near-real-time incremental replication that keeps destinations current while reducing the need for full reloads and limiting impact on operational systems. The new 2FA and SAML SSO options support integration with identity providers such as Okta, Azure AD, and Google Workspace, helping organizations align pipeline access with internal security and compliance policies.

CData Arc’s Q2 2025 release expands EDI processing, API integration, workflow customization, governance, and European e-invoicing capabilities. New metadata-aware routing through the Branch Connector, automated duplicate checking, and Workspace Send and Receive connectors improve EDI visibility and coordination between trading partner management and document workflows. Integration enhancements include Swagger imports, authentication configuration, and testing for REST APIs, plus connectors for DB2 on AS400/iSeries, Salesforce and Kintone batch APIs, branded web forms, ZPL label creation, PEPPOL, encoding transformations, and expanded ZUGFeRD support for multiple PDF attachments. Python scripting is now available in Script, XML Map, and Events connectors for custom logic and transformations, while security improvements add certificate-policy RBAC, SAML 2.0 support, certificate serial-number searches, and Vault API endpoints. On-premises users can download the release, while Arc Cloud users must contact support to arrange an upgrade.

CData Sync’s Q2 2025 release adds open mirroring for Microsoft Fabric, enabling near-real-time incremental replication of source data into Fabric-managed Delta Lake tables in OneLake. The new capability complements existing file-based delivery of Parquet, Avro, and CSV files, allowing teams to select ingestion approaches based on requirements for data freshness, latency, cost, and format. Mirroring is positioned for live dashboards and real-time analytics through change data capture or timestamp- and integer-based updates, while file delivery supports batch-oriented use cases such as archival, data science, and downstream transformations. By combining both methods in one platform, CData Sync aims to reduce the need for separate integration tools while supporting a range of Fabric data workloads.

CData’s Vibe Querying episode demonstrates how product leaders can use Connect AI’s remote Model Context Protocol server with Google BigQuery and Claude to analyze product telemetry through natural-language questions. Using CData’s SQL-based connector layer, which supports more than 350 data sources, the demonstration examines how AI can identify product-led growth “aha” and habit moments, such as users successfully retrieving data within 12 hours of signup and later running queries across multiple connections over several weeks. Claude also segmented successful users by department, identified high-performing connectors and client preferences, recommended onboarding and habit-building improvements, and drafted persona-specific email campaigns based on behavioral data. The example argues that conversational access to live business data can reduce analysis that traditionally requires months of cross-functional meetings, dashboards, and consulting work to minutes, while maintaining visibility into the generated SQL and using standardized, enterprise-oriented data access.

Enterprises choosing AI-agent architectures can use Model Context Protocol (MCP) to connect one centralized agent to multiple live systems or Agent-to-Agent (A2A) protocols to coordinate multiple specialized agents across bounded domains. MCP provides standardized, context-aware access to APIs, databases, and SaaS tools, making it suited to unified dashboards, monitoring, cross-functional automation, and operational assistance, though its central agent can become complex, a scaling bottleneck, and a single point of failure. A2A resembles a microservices model in which independently deployable agents collaborate through capability metadata and contextual messaging, supporting domain specialization, fault isolation, asynchronous workflows, and incremental scaling, but introducing challenges around coordination, shared context, security, governance, and accumulated latency. Rather than competing approaches, MCP is described as vertical integration between agents and enterprise applications, while A2A enables horizontal coordination among agent services; hybrid implementations can use MCP-connected agents for live data and tools while using A2A for broader workflows. MCP currently has stronger enterprise maturity, whereas A2A standards remain early and fragmented. CData MCP Servers are presented as a connectivity layer that abstracts APIs, authentication, and query translation so agents can securely retrieve real-time enterprise data and execute actions across systems, including in A2A-orchestrated multi-agent workflows.

CData Connect AI is presented as a tool for addressing schema mismatches, such as inconsistent field names, data types, and formats, that can complicate reporting and analytics and often require manual cleanup or scripting. It connects directly to SaaS applications, databases, and other sources using standard credentials, then enables users to shape datasets through a no-code Query Builder, a natural-language AI SQL Generator, or a SQL Editor for custom queries. Users can save configured queries as reusable, live derived views or virtual datasets that remain updated as source data changes. These prepared datasets can be connected to BI platforms including Tableau, Power BI, and Looker, with the aim of reducing post-extraction rework, shortening time to insight, and allowing business users to manage schema alignment independently.

Gartner Data & Analytics Summit sessions presented data fabric and data mesh as complementary approaches for modernizing data architecture rather than competing alternatives. Data fabric is described as an architectural layer that connects existing warehouses, lakes, cloud systems, and applications through active metadata and automation, improving access, governance, observability, performance, and adaptability without requiring a full technology replacement. Data mesh is framed as an organizational operating model that decentralizes ownership by having domain teams manage data as products, supported by self-service infrastructure and federated governance. Speakers emphasized that successful mesh adoption requires incremental implementation, cross-functional teams, product-oriented thinking, and cultural change. Although only 13% of surveyed organizations use both approaches together, Gartner reports that these organizations achieve stronger data product delivery outcomes, while metadata maturity remains a major barrier. Together, fabric supplies the technical foundation for intelligent connectivity and automation, while mesh distributes responsibility and aligns data delivery with business needs.

Agentic AI promises to complete business tasks across systems such as CRMs, ERPs, databases, and cloud applications, but its effectiveness depends on secure, reliable access to distributed enterprise data. The proposed architecture uses SQL as a common language because large language models are extensively familiar with its standardized syntax, making SQL generation generally more practical than handling numerous proprietary APIs. CData connectors translate SQL queries and commands into the API calls required by more than 300 business systems, while the Model Context Protocol (MCP) provides a secure channel through which AI agents can access those connectors. This approach supports both retrieving live data and performing actions such as creating, updating, or deleting records, while applying the user’s existing permissions, enabling validation before changes, and producing auditable records of activity. By placing the AI agent as the primary interface and using CData and MCP to manage system-specific integrations, the model aims to simplify development, preserve security controls, and enable agents to act across an organization’s software environment.

CData has introduced drivers and connectors for Okta that provide standards-based, live SQL access to identity and access management data from reporting tools, databases, AI applications, and custom software. The offerings support JDBC, ODBC, ADO.NET, Excel, Tableau, Power BI, Python, and PowerShell, with planned options for data replication and cloud-to-cloud connectivity, while using OAuth authentication and SQL-92 querying across more than 200 supported tools. Through Okta’s Admin Management REST API, users can access and update data including users, groups, applications, system logs, policies, and devices, enabling analysis of access controls, audit events, and security activity. A Power BI Desktop example describes installing the connector, configuring an OAuth-enabled DSN after creating a custom Okta OAuth application, selecting the CData Okta source in Power BI, and building visualizations from the retrieved data.

Salesforce’s acquisition of Informatica is presented as increasing uncertainty for PowerCenter users, particularly around support timelines, product roadmaps, pricing, and the potential prioritization of SaaS-based offerings over on-premises or hybrid deployments. Organizations operating across multiple clouds, legacy systems, and regulated environments may face challenges if Informatica becomes more tightly integrated into the Salesforce ecosystem. The passage positions CData Sync as an alternative for teams seeking to migrate from PowerCenter gradually while retaining deployment control, hybrid architecture support, change data capture, broad system connectivity, and connection-based pricing without usage fees. It encourages organizations to evaluate CData Sync as a self-hosted, cloud-managed, or fully hosted option rather than waiting for further changes following the acquisition.

CData’s MCP Servers are presented as a connectivity layer for AI agents that allows them to access live data and perform actions across more than 350 enterprise applications without centralizing or replicating data. Using an agent framework consisting of roles, data and context, actions, channels, and guardrails, the approach focuses primarily on extending agents’ access to external systems while leaving their business role and user interface to the host platform. CData connectors translate SQL generated by large language models into source-specific API calls, enabling agents to query systems such as Salesforce, NetSuite, Adobe Campaign, Zendesk, Jira, and ServiceNow and to update records or trigger workflows across them. The model is designed to preserve source-system context, user permissions, governance controls, and audit trails, while using SQL as a common interface intended to simplify multi-system operations and avoid custom integrations or data preloading.

Zapier’s Salesforce MCP Server and CData’s Salesforce MCP Server represent contrasting approaches to AI data integration: Zapier provides roughly 22 preconfigured, action-specific tools designed for efficient and guarded execution of common workflows, while CData exposes full data models, relationships, APIs, and stored procedures through a standardized SQL-like layer. Tests found that both systems can perform basic CRM tasks such as extracting invoice information from PDFs, creating leads, converting leads, and retrieving records, but they differ in token use and flexibility. Zapier generally completes known actions with fewer API calls, although its predefined tools may produce unintended results in unusual scenarios, while CData requires more exploratory queries but offers more systematic access, supports custom fields more clearly, and can discover native platform capabilities such as Salesforce’s convertLead procedure. The comparison suggests that enterprises may benefit most from combining Zapier’s streamlined tools for frequent, standardized workflows with CData’s broader capabilities for data exploration, complex analysis, cross-system investigation, and deeper integrations.

Sage’s first annual conference in Atlanta highlighted how AI, automation, and cloud migration are reshaping accounting software and finance operations in 2025. While spreadsheets remain widely used because of their familiarity, vendors are promoting synchronized, real-time spreadsheet tools that connect directly to Sage and other data sources. The discussion emphasized that the quality and depth of individual data connectors can be more valuable than broad but limited connectivity, particularly for users needing access to specific Sage Intacct tables and detailed data models. It also noted that software providers increasingly face high costs and complexity when building and maintaining their own integrations, creating demand for embedded connector platforms. Across Sage products, finance teams are seeking real-time analytics and automated data processes to move beyond compliance and reconciliation work toward more strategic advisory roles.

Insights from the Gartner Data & Analytics Summit in London emphasized that modern data integration should be driven by measurable business outcomes rather than technology alone. Speakers advocated for cross-functional fusion teams that share responsibility for integration goals, rapid pipeline prototyping to validate value before scaling, and automation to improve platform maturity. Data products were distinguished from ordinary datasets by their curation, reusability, metadata, lifecycle ownership, service levels, and demonstrable business impact, with categories ranging from essential utility products to business-transforming driver products. Generative AI was presented as an increasingly embedded integration capability that can enhance metadata, lineage, observability, data contracts, and deployment optimization, while making strong contextual metadata and cost management more important. Overall, the sessions portrayed integration as a collaborative, product-oriented, and AI-enabled discipline, and the author positioned CData Connect AI as a governed connectivity platform intended to support rapid access and prototyping across SaaS, cloud, and on-premises systems.

Organizations in regulated sectors often need hybrid data architectures that combine the security of on-premises systems with Databricks’ cloud-based scalability and analytics capabilities. Integrating legacy platforms such as SAP, Microsoft Dynamics, IBM DB2, and SQL Server can be difficult because of proprietary structures, limited connectivity, agent-based infrastructure requirements, performance and change data capture needs, compatibility with existing reporting tools, and compliance obligations under standards such as HIPAA and GDPR. CData’s Databricks Integration Accelerator is presented as an agentless, no-code integration option that supports secure live data replication from complex on-premises sources to Databricks without local agents or VPN complexity. Its features include SOC 2 Type II security, support for customized schemas and nested data, CDC capabilities, and predictable connector-based pricing rather than consumption-based costs. Implementation through CData Sync involves configuring source and Databricks destination connections, creating replication jobs, selecting tables, and scheduling and monitoring data movement.

Power BI drillthrough enables users to move from high-level dashboard visuals to detailed report pages filtered by a selected value, helping them investigate underlying causes without rebuilding reports. Its usefulness depends on having comprehensive, current source data, since stale extracts can produce incomplete or misleading insights. CData Connect Cloud is presented as a no- or low-code platform that provides live, governed connections between Power BI and systems such as ERP, CRM, SaaS, and custom databases, while supporting authentication, encryption, row-level security, centralized credential management, and usage monitoring. It also allows organizations to create reusable virtual datasets that combine sources such as NetSuite and Salesforce, centralizing join and filtering logic so updates are reflected across Power BI and other tools. The proposed result is faster access to transaction-level or operational details, more consistent reporting, reduced reliance on exports and synchronization jobs, and greater trust in business intelligence decisions.

Snowflake provides scalable, low-latency cloud analytics and AI capabilities, but its effectiveness depends on integrating operational data from systems such as SAP, Salesforce, Microsoft Dynamics, and Workday, whose differing APIs and schemas can create silos, costly custom development, brittle ETL pipelines, and governance challenges. CData positions its low-code and no-code connectivity tools as a way to address these issues through prebuilt connectors for more than 270 systems, automated schema management, no-code pipeline design, live data federation, and integration with Snowflake’s security controls. The proposed approach aims to shorten implementation timelines, reduce maintenance costs, improve consistency across platforms, and give teams greater flexibility to support real-time dashboards, cross-functional analytics, and evolving data needs.

At Gartner’s Data and Analytics Summit in London, analysts emphasized that successful AI depends less on models alone than on preparing data, governance, infrastructure, and teams for specific real-world uses. Mark Beyer argued that AI-ready data is contextual rather than simply accurate or governed, requiring alignment with each model and use case, continuous monitoring for drift, and frameworks such as model cards and pipeline-based readiness checks. Sue Waite highlighted that many AI projects fail before production because of scattered, inaccessible, poor-quality, or insufficiently governed data, advocating active metadata, data-quality capabilities, and observability tools to provide visibility into data, pipelines, lineage, and compliance. Afraz Jaffri focused on the operational challenge of moving models from pilots to production, comparing effective AI engineering to a Formula 1 pit crew that relies on coordinated cross-functional teams, modular pipelines, automation, agile practices, and measurement of lifecycle bottlenecks. Together, the sessions presented AI adoption as an organizational and data-engineering discipline requiring intentional preparation, shared context, and continuous operational improvement.

CData positions its Snowflake Integration Accelerator as a way for organizations to reduce the manual effort, delays, and maintenance associated with ingesting data from expanding cloud, on-premises, and hybrid application environments into Snowflake. The post argues that brittle legacy pipelines, schema changes, and point-to-point integrations can slow analytics and AI initiatives, citing biotechnology and healthcare customers that faced extensive manual mapping and long pipeline runtimes. Its proposed solution combines batch and change-data-capture ingestion from more than 270 sources, live data access for applications such as Microsoft tools and Salesforce, and support for structured and unstructured data through Snowflake Cortex AI capabilities. CData highlights native compatibility with Snowflake technologies, broad enterprise connectivity, hybrid deployment options, and volume-independent pricing, while claiming that customers such as Empower Services reduced pipeline runtimes from 90 minutes to 8 minutes and automated much of their ingestion work.

Databricks can centralize and analyze data effectively, but its lack of native connections to financial systems such as ERP, CRM, and accounting platforms can leave FP&A teams reliant on siloed data, manual exports, and delayed insights. The passage argues that current financial planning, forecasting, budgeting, and performance tracking require complete, frequently updated data from systems including SAP, Workday, Salesforce, QuickBooks, and NetSuite. It presents CData Integration Accelerator for Databricks as a solution offering connections to more than 350 sources, change data capture for near-real-time analytics, cross-system modeling and joins, security controls including encryption and audit logging, and predictable pricing. CData Sync can replicate source data into Databricks through scheduled or real-time ETL jobs, while CData JDBC drivers support live source queries through the Databricks Spark engine, allowing organizations to choose between loaded and directly accessed data for financial analysis.

Marketing teams often struggle to unify data from platforms such as Google Analytics, Google Ads, LinkedIn, Salesforce Marketing Cloud, and HubSpot in Databricks because of incompatible data models, frequent source changes, demand for near-real-time reporting, and the need for custom API-based integrations. CData positions its Databricks Integration Accelerator as a solution through prebuilt connectors for more than 270 systems, no-code ingestion, schema evolution, incremental replication, metadata management, and error detection. The platform supports both ETL workflows through CData Sync, which loads source data into Databricks, and live access through JDBC drivers that query source systems using the Databricks Spark engine. CData cites customer results including NJM Insurance’s reported 90% reduction in integration build time, 66% lower operational costs, and faster ingestion, while also describing support for large-scale, append-only pipelines compatible with Databricks Delta Live Tables and Unity Catalog governance.

Customer 360 initiatives aim to unify sales, marketing, service, and financial interactions into a complete customer view, but organizations often face difficulties connecting diverse business systems, meeting real-time data needs, conserving engineering resources, and maintaining governance. CData positions its Databricks Integration Accelerator as a no-code solution for these challenges, claiming it can deliver Customer 360 insights faster and at lower cost than legacy integration approaches. The platform provides connectors for more than 270 enterprise sources, supports both ETL-based loading into Databricks and live access through Lakehouse Federation, and includes incremental synchronization, change data capture, and schema-evolution capabilities. It also integrates with Databricks Unity Catalog and offers role-based access controls and audit logging to support privacy, lineage, and compliance requirements. Implementation involves connecting source systems, configuring Databricks as a destination, and creating scheduled incremental ETL jobs, while the broader goal is to let organizations improve data freshness, reduce technical bottlenecks, and expand governed access to customer insights.

CData Connect AI is presented as a data virtualization platform designed to help organizations protect personally identifiable information, including names, addresses, and social security numbers, amid growing security and privacy risks. PII obfuscation through masking, encryption, anonymization, and pseudonymization can reduce unauthorized access, support compliance with regulations such as GDPR and CCPA, preserve data usefulness for analytics, and strengthen customer trust. The platform provides Derived Views for custom transformations, Workspaces for centralized querying and governance, and user-defined credentials that maintain individual identity and role-based access controls. Its security capabilities include configurable data masking with realistic fictional values, encryption for data in transit and at rest, granular access management, and anonymization methods such as shuffling and generalization, allowing organizations to protect sensitive data while integrating it across diverse data sources and workflows.

AI forecasts can fail despite accurate inputs and sound models when they lack real-time context such as shipping disruptions, product defects, weather events, inventory outages, or staffing changes. Traditional AI typically relies on static or manually refreshed data, which can leave important operational signals unavailable and produce decisions based on outdated snapshots. Context-aware AI addresses this gap by incorporating live sales data, operational metadata, and field updates, enabling forecasts and recommendations that better reflect current conditions. The Model Context Protocol (MCP) is presented as a standard for delivering relevant, up-to-date information into AI systems, helping data teams shift from correcting inaccurate outputs to interpreting trends, identifying anomalies, and supporting timely decisions. CData MCP Servers are described as a way to connect fragmented data sources and provide AI tools with real-time context across systems and teams.