January 2026 Summaries

Organizations exploring data-intensive opportunities, such as AI, often need to reevaluate their data stacks to support modern workloads, making the concept of total cost of ownership (TCO) crucial for understanding the long-term expenses of data management. Fivetran offers a TCO calculator, a self-serve tool designed to estimate the current costs of building and maintaining data pipelines, comparing them with the efficiencies of Fivetran's managed services. The tool allows users to input their company size and adjust key factors like engineering time, maintenance, and downtime to gain insights into hidden costs often overlooked in budgets. The calculator highlights the impact of downtime, which can significantly affect productivity and decision-making, and provides default estimates based on industry benchmarks to guide users in understanding their current and potential future expenses. This enables organizations to better assess the trade-offs between DIY pipelines and managed services, potentially reducing operational burdens and ownership costs over time.

Salesforce enhances customer relationship management by integrating with Fivetran Activations, enabling businesses to leverage enriched, actionable data for AI-powered insights. This integration allows for the creation of AI Columns, which generate context-aware insights directly within Salesforce. Use cases such as automating fit scoring and building AI product usage summaries demonstrate the potential for transforming CRM workflows. Automating fit scoring involves setting up an Ideal Customer Profile (ICP) analysis, crafting GPT prompts for scoring based on customer data, and testing to refine accuracy. Building AI product usage summaries involves consolidating product event data, analyzing trends with GPT, and structuring insights for Salesforce consumption. These processes empower go-to-market teams by providing actionable insights, ultimately guiding account executives toward high-value opportunities, preventing churn, and enhancing marketing campaigns. Through AI Columns, Salesforce teams can achieve improved efficiency and effectiveness, driving revenue growth and customer engagement.

One-Click Audiences is a feature designed to streamline the process of audience activation for marketing teams by allowing them to send their customer data to major advertising platforms like Facebook, Google, and LinkedIn with a single click. Developed by Census, now part of Fivetran, this feature automates the synchronization of audience data, maximizing ad match rates through comprehensive identifier mapping and ensuring proper data formatting, hashing, and normalization. By automating these processes, marketers can focus more on strategic tasks such as testing messaging and managing campaigns. The setup involves a one-time data definition process by data teams, who define person entities and configure identifier mappings, followed by marketers using the Audience Hub to create and activate audiences. This approach eliminates the need for manual CSV imports or SQL queries, allowing for more efficient and intelligent targeting across over 200 engagement channels.

The evolution from human-centered institutions to computational systems hinges on placing data at the core of operations, transforming it from a retrospective tool into an essential transaction medium for AI-driven workflows. Unlike traditional data usage, where humans relied on data for interpretation and decision-making, AI agents now require data to be immediately accessible, interpretable, and enforceable at the point of action, reshaping the data infrastructure to meet these needs. This shift involves redesigning data systems to enable natural language interfaces for broader access, establishing agents as universal task executors, and redefining operational systems as rule-governed environments. The data layer becomes the active environment for decision-making, necessitating high-quality, universally accessible data with clear definitions and rules to support agentic operations. As organizations transition to this AI-driven model, building a stable, interoperable data foundation becomes crucial, enabling AI systems to operate effectively within workflows and ensuring competitive advantage in the emerging agentic future.

The Fivetran Connector SDK enhances data teams' ability to create custom integrations with proprietary systems, internal APIs, and niche data sources by handling infrastructure, scheduling, and data delivery. The introduction of the "fivetran init" command simplifies the setup process by providing a scaffolded project structure, clear starter code, and pre-configured contexts for AI-assisted coding environments like Claude, Cursor, and VSCode with Copilot. This command streamlines development by offering template-based initialization and enabling AI assistants to generate code that adheres to Fivetran's best practices. It supports quick prototyping by reducing setup friction, thus facilitating faster and more efficient custom connector development. The SDK exemplifies Fivetran's dedication to extensibility, offering a streamlined way to build custom connectors for unique organizational needs alongside its library of over 700 pre-built connectors.

The text discusses the use of Claude Code, a CLI-first, editor-agnostic tool, in conjunction with the Fivetran Connector SDK to streamline the development of data connectors, particularly for the FDA Drug API. Claude Code enhances connector development through repository-aware AI, enabling it to understand and apply SDK examples and best practices, which allows developers to transform the traditionally lengthy engineering task into a collaborative 30-minute session. The process involves cloning a repository, prompting the AI with specific requirements, and iterating through testing and validation phases, ultimately deploying a production-ready connector. Claude Code's advanced reasoning capabilities and contextual awareness facilitate intelligent collaboration between developers and AI, shifting the developer's role towards solution architecture and validation, rather than writing every line of code. This approach enables rapid prototyping and production-grade solutions by leveraging AI for intelligent pattern matching and repository understanding.

Apache Polaris is an open-source catalog for Apache Iceberg tables, offering a standardized REST interface to manage metadata, which enables query engines to access tables without embedding technology-specific code. Fivetran employs Polaris in its Managed Data Lake Service, utilizing it as the default catalog for Iceberg tables, with support for destinations like Amazon S3 and Google Cloud Storage. The service ensures data integrity by allowing only Fivetran to modify catalog metadata, while users are advised to query tables through the catalog rather than directly accessing raw Parquet files to avoid issues like file discovery overhead and lack of ACID guarantees. Polaris offers benefits such as snapshot isolation, transactional consistency, and seamless schema evolution, which are crucial for maintaining data integrity and performance. Various query engines, including Apache Spark, Snowflake, and Trino, are compatible with Polaris, with OAuth2 authentication facilitating secure access. As Polaris evolves, it is expected to graduate from the Apache Incubator by late 2025, with improvements in DuckDB and Snowflake support anticipated.

Fivetran has developed an internal tool to manage the overwhelming number of feature requests for its over 700 connectors by integrating data from various sources with AI agents, creating a customer feedback intelligence platform. This system utilizes Fivetran connectors, BigQuery, dbt, and Census to centralize data, which is then transformed into a unified model to provide a complete picture of each request. The tool employs AI to triage requests, generate feasibility reports, and draft personalized customer communications, significantly reducing the cognitive load on product managers and enhancing response times. The AI-driven approach improves prioritization, closes feedback loops, and maintains a human touch in customer interactions, while also offering lessons on the importance of clean data and aligning technology with workflow needs. This innovation not only streamlines internal operations but also enhances customer satisfaction and trust.