Fivetran acquired Census to support our mission of making reliable data easily accessible across the entire organization, regardless of technical acumen. Together, our ease of use and AI functionality abstract complexity to make an analytics workspace for all. Within our joint offering, users of all skill levels can now easily generate new datasets and transform data using natural language — enabling fast, independent access to the right data without delays or manual query building.

Technical barriers, such as a lack of SQL fluency, often hinder business agility and prevent stakeholders from independently exploring data and generating insights. Even technically savvy users may struggle to navigate undocumented schemas and constantly evolving datasets, forced to rely on tacit knowledge or constant clarification from engineering and analytics teams to write accurate queries. This cycle of dependency delays insights, limits experimentation, and ultimately constrains the organization’s ability to respond to change.

True business agility requires stakeholders to become data superusers — empowered to ask questions, explore data, and generate answers independently, without waiting in line for technical support.

From question to data product: no code, just language

Once Fivetran has loaded your data, Census makes it easier than ever to use your data to take action. AI-assisted SQL generation enables users to use natural language prompts, like “All accounts with open opportunities,” and automatically generate the underlying SQL to create actionable insights that can be shared and scaled across the business. 

Critically, this feature does not send your data to an external black-box AI model. Instead, it uses your existing API keys for providers like Gemini, Anthropic, or OpenAI, ensuring your data remains governed by your current usage terms. Only dataset names and column metadata are shared to enable context-aware SQL generation — your business-sensitive data never leaves your environment.

The result? Analysts save hours of manual exploration. Datasets can quickly be pushed to downstream applications, allowing business users to decrease time-to-insight and self-serve faster. IT teams maintain control over governance and lineage. Your business teams know the data they need in their operational workflows and can now generate insights without deep SQL knowledge.

No code, augmented data transformation

After generating your initial dataset, AI Columns seamlessly enrich your data with intelligent, context-aware transformations. Whether it's auto-labeling sales opportunities by intent, summarizing call notes into one-liners, or inferring missing job titles from LinkedIn data, you can enhance your data products with powerful, reusable AI-driven logic.

AI Columns make categorizing large and complex datasets easier and more scalable by leveraging large language models (LLMs) like OpenAI, Claude, and Gemini. Categorization is a powerful technique for enriching and transforming data — whether it's structuring user feedback or evaluating leads using firmographic attributes—but doing this manually can be time-consuming and error-prone. AI Columns streamline this process by allowing users to define categories, craft structured prompts, and apply consistent categorization rules at scale. 

Importantly, all AI operations maintain enterprise-grade control: logic runs within the warehouse, and only metadata is shared with models, ensuring strict data privacy and compliance. Admins can also disable AI features at any time to align with internal governance and policy requirements.

Natural language interfaces enable data superusers

AI advancements like Fivetran’s natural language interface for SQL generation accelerate technical teams’ work while enabling humans to ask intelligent questions and get actionable data back, creating an entire organization powered by data superusers

As AI and data analytics continue to grow in capability, Fivetran ensures you unlock that potential securely, intelligently, and at enterprise scale.

[CTA_MODULE]