Introducing: AI Assistant, The First Chapter of Secoda AI

Secoda AI is the first AI suite of search, catalog, lineage, and documentation solutions, powered by LLM
Last updated
May 2, 2024

Update: Secoda AI is live! See it in action here

Secoda AI - the first AI suite of search, catalog, lineage, and documentation solutions, powered by LLM.

Meet AI Assistant, the first member of the Secoda AI suite

At Secoda, we believe the future of data discovery, just like the future of search, will be driven by large language models that allow anyone to ask any question to their data. Our mission has always been to give everyone the confidence to use data - this includes both technical and non-technical business teams. In order to do that, we want to make the experience of understanding and using data as fast and contextual as using natural language search. That’s why we’re extremely excited to launch the Secoda AI Assistant, the first member of Secoda AI - our suite of AI powered solutions to enhance your data workspace.

The future of data discovery

One of the definitive aspects of AI advancements in recent months is how quickly new models are developed to handle an ever growing range of use cases. And they’re good. Each advancement is moving the world of AI from hype to reality at an unbelievable pace.

We see the new frontier of data discovery beginning now. Typically, that process has been layered and complex. Knowledge is generally widely spread across various databases, warehouses, transformation tools, BI dashboards or held in the minds of a select few people.

The potential is endless and we’re so excited to be part of the movement that is shaping the future of how companies find, use, and consume data.

AI is going to increasingly help data and business teams easily search for answers by asking questions like:

  • What tables have been transformed in dbt?
  • What are tables with the least amount of documentation?
  • What is the best dashboard to look at revenue?
  • What tables have been recently updated?
  • What does MRR mean?

However, this level of data discovery will depend extensively on well-indexed and maintained metadata so data practitioners and consumers alike can scan expansive databases and subscribe to content that they find useful.

Why we made Secoda AI: natural language search to change the data catalog landscape

We believe data catalogs in the market today are fundamentally flawed. They are are clunky and painful to implement - and even more burdensome to maintain. On top of this, they are difficult and inaccessible to non-technical users who are the consumers of most data products. The result is that data catalogs are only adopted and used within the data team and do not solve the issue of democratizing data to reduce the burden on data teams by reducing the volume of requests they receive from business teams.

At Secoda, we view a data catalog as a feature within an organization’s data workspace. As we continue building the homepage for your data, we believe that to truly enable data discovery and adoption of data products, the data workspace must be easy for data teams to update and maintain while being just as easy to access and consume for business stakeholders.

  • Secoda AI for data teams: Let’s be honest, documentation is not everyone’s favorite task. It’s often tedious, time consuming, and requires a lot of manual effort to create and maintain both accuracy and completeness of the documentation, which can lead to frustration and lack of motivation. Secoda AI was built to help solve this through automating hours of manual work so data teams can spend less time documenting and more time on impactful, value creating work.
  • Secoda AI for business stakeholders: Natural language search is at the core of how we are creating he future of data adoption. To truly bridge the gap between data practitioners and non-technical consumers, a business user will be able to submit a natural language question and get reliable answers, context, and direction from the data workspace.

What are the current functions of AI Assistant?

Turn Text into SQL

Based on your existing column and table definitions, Assistant AI will turn natural language into SQL allowing you to write queries faster on new tables and dashboards.

Automatically generate documentation

Using existing table descriptions, column descriptions and dictionary terms, AI Assistant will create documents surrounding the context of each asset so you can focus more of your time on creating business value.

Automatically tag PII data

AI Assistant surfaces potential PII data from across your data stack and streamlines your data governance workflows.

What are the current limitations of AI Assistant?

All AI models, including AI Assistant, face certain limitations which are are actively being improved. Reasons for these limitations include narrow data inputs, prejudice in the system, not enough training or testing data, or poor quality of the data used. As a consequence, AI Assistant may output incorrect information.

Despite these limitations, AI generated recommendations and machine learning models will continue to improve over time as more data is collected and algorithms are refined. Secoda’s AI Assistant is the same - this is only our beta version. As we improve our models and address additional, specific use cases, this feature will only improve in the months to come.

AI continues to be a prioritized area of innovation for Secoda and a major focal point of our product roadmap. With more data and better algorithms, AI generated recommendations will become more accurate and more effective at providing personalized recommendations to users.

Try Secoda for Free

Learn more about Secoda AI and see it in action here

Keep reading

See all stories