Select columns bring structure to your Phacet by letting you classify input data through a fixed list of predefined tags applied to each record.

• Single Select: choose one tag from the list.
• Multi Select: celect multiple tags for a single record.

Combined with AI or Python, they unlock automation workflows and give you powerful tools to triage and structure information at scale — with precision and consistency

Example use cases

• Classify inbound requests: assign issue type, urgency, or department to support tickets and contact forms.
• Analyze qualitative feedback: apply sentiment or theme tags to open-text survey responses or reviews.
• Label financial documents: tag contracts, invoices, or reports by document type, risk level, or business unit.
• Triage legal clauses: Flag clauses by category (termination, liability, confidentiality) in extracted contract data.

Setting Up a Select Column

Step 1: Create the column

• Add a new column to your Phacet.
• Choose Single Select or Multi Select as the column type.
• Give it a clear, descriptive name (e.g. Request Type, Tags, or Theme).

Step 2: Define your tags

• Enter the list of allowed tag values in the column settings.
• Use concise, unambiguous tag names
• Optionally, assign a unique color to each tag for fast visual parsing

Step 3: Select your classification tool

Decide how the tag values will be applied:

• Manual Input: for human review and tagging.
• AI (LLM): AI will analyze input data and assign one or more tags automatically.
• Python Tool: if your classification relies on rule-based or structured logic.

Let’s see below how to select the best tool for your workflow

Manual vs Automated tagging

Select columns support manual input or automated generation.

Both Python and AI tools can populate your select columns automatically. Select the right method depending on your classification method:

Once you've selected the tool for your Select column, here’s how to configure it properly.

Referencing other data

When using the AI Tool or Python Tool, you can reference data from other columns using the @ symbol followed by the column name (e.g. @Feedback, @Document Text).

This allows your tool to use the value from that column as input for classification.

AI Tool setup

• The AI tool can interpret the meaning of tags defined in the column settings and will automatically assign the most relevant tag(s).
• You don’t need to repeat the tags or their definitions in your AI instructions—unless you want to guide the classification logic more explicitly.
• By default, AI can return no tag if it considers that no tag match the requested classification. If you want to force a tag to be returned, you must specify that condition in your AI instructions.

Lear more about the AI tool in the dedicated article

Python Tool setup

• Single Select: your Python function must return a string that exactly matches one of the predefined tags in your column settings.
• Multi Select: your function should return either an empty list (no tag) or a list of valid tags from your column settings.

Lear more about the Python tool in the dedicated article

Example prompts and scripts

AI Tool

Categorize a GDPR request (Single Select)

Read the incoming request in @RequestText and categorize it as one of the listed tags.

Multi-tag feedback themes (Multi Select)

Read the comment in @FeedbackText and select all relevant tags from the list defined in this column.

💡 Make sure your instructions clearly describe how tags should be selected. The AI will map outputs to your tag list automatically.

Python Tool

Client request categorisation (Single Select)

def main():
    if "refund" in @EmailBody():
        return "Billing Issue"
    return "General Request"

Product feedback categorisation (Multiple Select)

def main():
    text = @Comment
    tags = []

    if "price" in text.lower():
        tags.append("Pricing")
    if "slow" in text.lower():
        tags.append("Speed")
    if "support" in text.lower():
        tags.append("Support")

    return tags

The value returned must match one of the tags defined in your column settings.

Best practices

• Keep tag lists short and clear: Avoid ambiguity and overlap between options.
• Use consistent naming: Ensure tag names are explicit and mutually exclusive.
• Color-code for readability: Helps users scan records quickly.
• Validate with AI or Python: Combine with logic or prompts to maintain quality.

Select columns bring structure to your Phacet. Combined with AI and Python, they give you powerful tools to classify, triage, and structure information at scale — with precision and consistency.