Columns

Columns are where the actual work happens in Cotera. While datasets provide the foundation and agents bring the intelligence, columns are where that intelligence gets applied to your data to create tangible business value.

What Are Columns?

Columns represent different types of work that can be performed on your data within a dataset. Every column takes your existing business data and transforms it, analyzes it, or uses it to trigger actions.

When you add a column to a dataset, you're essentially adding a new capability that processes every row of your data. The results appear as a new column alongside your original data, creating an enriched dataset that combines raw information with intelligent outputs.

How Columns Connect the Foundation

Building on Datasets

Remember that datasets serve as your AI foundation by connecting to your data warehouse and providing the workspace where intelligence operates. Columns are how you actually deploy that intelligence.

Each dataset can contain multiple columns, each performing different types of work on the same underlying data. This means you can layer multiple types of analysis and automation on top of a single data source.

Agents in Action

The agents we discussed earlier don't exist in abstract - they live and operate as specific types of columns within your datasets. When you create an agent to analyze customer sentiment or categorize support tickets, you're actually creating an agent column that applies that intelligence to every row in your dataset.

Integrated Intelligence

This column-based approach means AI insights don't live in isolation. Agent outputs become part of your data structure, available for reporting, analysis, and feeding into other business systems just like any other data field.

Types of Columns

Cotera supports several types of columns, each designed for different kinds of work:

Agent Columns

These deploy AI agents to analyze and interpret your data. Agent columns can read text, process images, make decisions, and generate insights based on the business logic you define.

Formula Columns

Similar to spreadsheet formulas, these perform calculations and data transformations using your existing data. Formula columns handle mathematical operations, data cleaning, and logical operations without requiring AI.

Tool Columns

These connect directly to external tools and services, allowing you to send data to other systems or fetch additional information. Tool columns can update CRM records, send notifications, or pull data from APIs.

Taxonomy Columns

These provide a user-friendly interface for categorizing data into predefined topics or themes. Taxonomy columns let you build sophisticated classification systems through an intuitive UI, making it easy to organize content into categories without writing complex conditional logic.

Each type serves different purposes, and you can mix and match them within the same dataset to create comprehensive automated workflows.

The Column Workflow

1. Define the Work

Decide what you want to accomplish with your data - analysis, transformation, or action.

2. Choose Your Column Type

Select the appropriate column type based on your needs: agent for intelligence, formula for calculations, or tool for integrations.

3. Configure the Logic

Set up the specific instructions, formulas, or tool connections that define how the column should process your data.

4. Deploy and Monitor

The column processes every row in your dataset, adding its outputs as a new field in your data structure.

Why the Column Approach Works

Scalable Intelligence

Once configured, a column applies its logic to every row in your dataset automatically. Whether you have hundreds or millions of records, the same intelligent processing scales seamlessly.

Integrated Results

Column outputs become part of your standard data infrastructure, making AI insights immediately available for business intelligence, reporting, and decision-making.

Flexible Composition

You can combine different column types to create sophisticated workflows. An agent column might analyze customer feedback, a formula column could calculate risk scores based on that analysis, and a tool column could send alerts for high-risk situations.

Iterative Development

Columns can be modified, tested, and refined without disrupting your underlying data or other columns, making it easy to improve your automated processes over time.

From Data to Action

Columns transform static business data into dynamic, intelligent operations. Your customer database becomes a customer intelligence system. Your transaction records become automated financial analysis. Your support tickets become proactive issue detection and response.

This is where Cotera's promise of "AI agents that do things, not just say things" comes to life - through columns that don't just analyze your data, but actively work with it to drive business outcomes.

The column is where your data meets AI capability, creating automated intelligence that scales with your business operations and integrates seamlessly into your existing workflows.