AI enrichment lets you add intelligence to your data by running AI-generated analysis on each row of your table. Instead of just looking up facts, AI columns can create personalized messages, score leads, summarize companies, classify prospects, and more.
What AI enrichment can do
- Generate custom outreach messages using each lead's name, company, role, and other details.
- Analyze multiple data points to assign a fit score or priority level.
- Condense company descriptions into concise, actionable summaries.
- Categorize leads by industry, intent, persona, or any criteria you define.
- Evaluate the tone of social posts, bios, or other text fields.
- Pull structured information from unstructured text fields.
Setting up an AI column
Step 1: Create the column
In your table, click Add Column and select AI Column as the column type.
Step 2: Choose an AI model
Select from the available AI providers. Different providers vary in quality, speed, and cost:
- Some are better at creative writing and nuanced language tasks.
- Others are faster and cheaper for straightforward classification or extraction.
- Consider starting with a mid-tier option and upgrading if you need higher quality output.
Your available models depend on your TexAu plan. Premium plans unlock access to more advanced models.
Step 3: Write your prompt
Write the instructions that tell the AI what to do with each row's data. Use column variables (see below) to reference values from other columns.
Step 4: Test
Before running on your full dataset, test the prompt on a few rows. Review the output carefully:
- Is the format consistent?
- Is the quality acceptable?
- Are the column variables pulling in the right data?
Adjust your prompt based on the test results.
Step 5: Run
Once you're satisfied with the test output, run the AI enrichment on your selected rows or full dataset.
Using column variables
Column variables let you insert data from other columns into your AI prompt. Use double curly braces with the column name:
Write a personalized outreach email to {{firstName}} who works as {{jobTitle}} at {{companyName}}.
Mention their recent activity: {{linkedinHeadline}}.
Available variables match your table's column names exactly. If a column is named "First Name", use {{First Name}}.
Make sure the columns you reference have data before running AI enrichment. Empty variables produce poor results.
Credit costs
AI enrichment uses credits based on:
- The AI provider and model selected (more capable models cost more).
- The length of your prompt and the expected output.
- The number of rows processed.
Check the estimated credit cost before running. Start with a small batch to verify the per-row cost matches your expectations.
Tips
- Be specific in your prompts. Vague instructions produce inconsistent results. Tell the AI exactly what format, tone, and length you want.
- Test on 5-10 rows first. Always validate output quality before processing hundreds of rows.
- Use the right model for the job. Simple classification tasks don't need the most advanced model. Save premium models for tasks that require nuance.
- Keep prompts focused. One task per AI column. If you need both a lead score and a personalized message, create two separate AI columns.
- Include enough context. The more relevant column variables you provide, the better the AI can personalize its output. But don't include irrelevant data that might confuse the model.