July 2026 Summaries

Automated data integration for IBM Z mainframes, particularly using Fivetran's Db2 for z/OS connector, offers a modern solution to accessing and analyzing mission-critical data stored in Db2 without the high costs and inefficiencies associated with traditional methods. Instead of running expensive and slow nightly data dumps or maintaining fragile hand-built ETL pipelines, Fivetran leverages change data capture technology to read transaction logs, capturing only data changes since the last sync, which reduces CPU usage and associated costs. This approach not only provides continuous, up-to-date data for analytics but also allows for significant savings by utilizing zIIP processors, which are not included in standard IBM software billing. Implementation is straightforward, requiring minimal setup and maintenance, and allows organizations like banks, insurance companies, and healthcare systems to perform real-time analytics and decision-making without disrupting mainframe operations.

Open Data Infrastructure is a concept advocated by Fivetran and dbt Labs, emphasizing the importance of interoperability and open standards in data management to ensure flexibility and adaptability to evolving technological demands. This approach supports storing data in open formats, allowing it to be utilized across various tools, compute engines, and AI systems without vendor lock-in, thus enabling organizations to scale operations effectively while managing costs. A modern data lake serves as the cornerstone of this infrastructure by decoupling storage and compute, preserving the optionality to choose components freely and adapt to future developments. In contrast, closed data infrastructures are characterized by limited data portability and high-risk commitments, often leading to increased engineering complexity and costs. Key challenges to maintaining an open infrastructure include avoiding architectural enclosure, vendor lock-in, and fragmented semantics, which can hinder data usability and governance. As AI advances, this openness becomes crucial for supporting machine-readable data and ensuring that data, metadata, and governance structures remain intact amid rapid changes in technology, allowing organizations to preserve their data assets while seamlessly integrating new systems and tools.

The Google Cloud Rapid Agent Hackathon showcased innovative AI agents that leverage real-time, trusted data to address diverse challenges across sectors such as genomic medicine, SaaS operations, and ecommerce. Among the 1,430 submissions, standout projects like "Unravel," "TruthKeeper," and "How Did I?" demonstrated the potential of AI agents to enhance decision-making by providing clinicians with updates on genetic research, detecting inconsistencies in business data, and offering operational insights for ecommerce businesses. These solutions highlighted the importance of reliable data in enabling AI agents to effectively support human expertise, improve operational efficiency, and prevent costly mistakes, emphasizing that the true value of AI lies in its ability to assist in making informed decisions rather than simply generating answers.

As artificial intelligence (AI) demands evolve, many existing data stacks struggle to support agentic AI, which requires real-time data, consistent semantics, and robust access paths. The Open Data Infrastructure (ODI) offers a solution by promoting an architectural shift away from proprietary, warehouse-centric stacks to open-standards-based infrastructure, facilitating seamless data integration and accessibility across tools and environments. ODI emphasizes three pillars: automated ingestion and transformation for flexibility, an open data lake to prevent vendor lock-in and reduce duplication, and a unified semantic layer to ensure consistent data understanding across systems. The ODI Assessment, an 8-question quiz, helps organizations evaluate their current data architecture's readiness for AI and identify areas for improvement, encouraging the adoption of an ODI framework that caters to both human and AI consumers efficiently.

Fivetran accelerates mergers and acquisitions (M&A) by providing a robust data infrastructure that facilitates each phase of the process, from initial market intelligence through to post-merger integration. The key to successful M&A lies in the quality, speed, and completeness of data access, which separates successful acquirers from those who may overpay or fail to realize expected synergies. Fivetran's capabilities, such as its Connector SDK and managed connector catalog, enable companies to ingest, normalize, and analyze data from multiple sources rapidly, allowing for proprietary insights and efficient due diligence. This infrastructure supports quick initial financial modeling, deep dive diligence across various functional domains, and seamless integration of data systems, thereby reducing risk and ensuring continuity post-acquisition. By providing a unified operational picture and enabling data activation in operational tools, Fivetran helps companies achieve faster, deeper diligence, predictable integration, and operational performance that aligns with their investment thesis, ultimately creating a competitive advantage in M&A transactions.

Sharing data across tools is crucial for organizations to ensure unified context, reduced friction, and flexibility in their data management processes, especially as they increasingly rely on AI and automation. Open Data Infrastructure (ODI) addresses the challenges posed by siloed data systems by providing a consistent, shared data foundation that prevents misaligned decisions, unreliable AI outputs, and costly infrastructure overhead. By eliminating duplicated data across environments, organizations can decrease compliance and governance issues, reduce human bottlenecks, and allow engineering teams to focus on innovation rather than maintenance. This approach also preserves flexibility by avoiding vendor lock-in, enabling organizations to adapt to evolving technologies without the need for complete system overhauls. Ultimately, a unified data strategy not only streamlines auditing processes but also prepares organizations to capitalize on AI advancements by ensuring all agents and systems operate from the same set of facts, thereby enhancing their competitive advantage.

Fivetran has expanded its support for IBM Db2 databases by introducing managed-service connectors for Db2 for z/OS and Db2 for LUW, completing its Db2 portfolio. These connectors are designed to overcome the unique challenges posed by each Db2 environment, enabling efficient data replication for analytics, AI, and reporting without overburdening the source systems. For Db2 on z/OS, Fivetran leverages the Db2 transaction log for data replication, reducing on-mainframe workload and offering significant cost savings by utilizing zIIP processors for DRDA connectivity. Similarly, for Db2 on LUW, Fivetran uses the db2ReadLog API for transaction log reading, ensuring minimal impact on the source database. Both connectors allow organizations to manage their data pipelines through Fivetran's platform, offering continuous data updates for various analytics platforms without the need for complex custom solutions or legacy replication tools.

Fivetran's Salesforce connector offers a streamlined solution for transferring data from Salesforce to a data warehouse, mitigating the complexities typically associated with Salesforce's API, including handling formula fields, custom objects, and daily API limits. By utilizing OAuth for connection, the connector automatically discovers and syncs all accessible objects and fields, including custom ones, without requiring prior specification, making it suitable for teams with varying technical expertise. Fivetran also accommodates unstructured data by replicating files like PDFs and images alongside structured records, ensuring complete data integration. Additionally, it manages API usage efficiently by switching between Bulk and REST APIs based on the data sync requirements, thus preventing disruptions due to exceeded request limits. The integration of dbt Quickstart package translates Salesforce formula fields into SQL views, maintaining data integrity and providing seamless access to up-to-date information, ultimately allowing teams to focus on data analysis rather than extraction logistics.

Amazon's decision to retract its proposed fees for the Selling Partner API, initially announced in November 2025, highlights the power of community backlash in influencing corporate policy. The company had planned to implement a $1,400 annual subscription and usage fees, sparking immediate opposition from developers and agencies who argued that such costs would disproportionately affect small and mid-sized developers and ultimately burden millions of businesses reliant on Amazon. In response to the pushback, Amazon postponed and eventually canceled the fee rollout by May 2026, illustrating that collective advocacy can lead to significant changes. This incident underscores a broader trend where vendors, like SAP, Slack, and Workday, are increasingly restricting data access by imposing fees or terms, creating "walled gardens." However, Amazon's reversal demonstrates that sustained pressure can prompt vendors to reconsider these strategies, aligning with Fivetran's belief in seamless data mobility. This approach is critical as AI-driven processes, which make extensive API calls, become more prevalent, necessitating vendor accountability to ensure data accessibility without prohibitive costs.

Fivetran is enhancing its data integration capabilities to support the needs of modern AI systems by moving beyond traditional structured data pipelines to include unstructured and semi-structured data. While conventional pipelines focus on structured data like CRM and sales records, Fivetran now enables the ingestion and management of unstructured data, such as customer support tickets, email threads, and contract PDFs, which are vital for AI applications. This expansion allows data to be stored in cloud data warehouses like Snowflake, BigQuery, or Databricks, where they can be processed using native AI tools for tasks such as OCR, text extraction, and audio transcription. Fivetran's connectors facilitate the transformation of data from a variety of sources, including files and SaaS platforms, into structured formats suitable for AI model training and operations, while dbt further helps in modeling and transforming this data for AI-driven insights. This comprehensive data management approach ensures that AI systems have access to rich, contextual data, thereby enhancing their utility and effectiveness.

Fivetran, a connectivity platform, offers five distinct paths to integrate virtually any data source into a destination, accommodating sources not listed in its extensive catalog of over 750 connectors. These paths include Managed Connectors for popular SaaS applications and ERPs, File Connectors for structured or unstructured file formats, Database Connectors for direct access via log-based change data capture, a Connector SDK for proprietary APIs and niche sources, and Webhook Connectors for real-time event data push. Each path is tailored to different source types and levels of engineering effort, ensuring comprehensive data integration capabilities for organizations regardless of their unique system requirements.

Retail and consumer packaged goods (CPG) teams in the Asia-Pacific (APAC) region are shifting their focus from merely centralizing customer data to actively utilizing it across operational systems to drive growth and competitive advantage. While centralization has laid the groundwork, the emphasis is now on data activation, which involves integrating trusted data into systems where employees, customers, and AI can leverage it for decision-making. This shift is driven by the need for personalized customer experiences amidst tighter margins and complex tech environments. Activation extends beyond marketing to enhance CRM, loyalty programs, and support services by providing richer context and more relevant customer interactions. The portability of data is crucial, as organizations aim to avoid vendor lock-in while maintaining architectural flexibility. Fivetran plays a pivotal role by enabling businesses to operationalize data across various systems, reducing manual processes and enhancing the value derived from existing data investments, thereby allowing companies to adapt to diverse markets and technologies.

As vendors increasingly impose controls on data access, enterprises face the challenge of navigating these restrictions while attempting to harness AI capabilities, with the concept of a "walled garden" emerging as a significant barrier. The difficulty lies in the unification and centralization of data due to various vendor-imposed limitations such as restrictive API access, expensive pricing models, and stringent terms of service, which collectively hinder the ability to replicate and integrate data across platforms. This issue becomes more pronounced as AI demands access to diverse and expansive data sets from multiple systems to function effectively, leading to a situation where data movement and access are more complex than the technical AI solutions themselves. As a result, organizations are encouraged to prioritize Open Data Infrastructure (ODI) in their software purchasing decisions to ensure long-term innovation and control over their data, allowing them to build and analyze without constantly renegotiating access.