Key Capabilities
Column Definition
Create custom fields on your territory account table. Define column types, names, and what they should contain. Use AI to suggest column structures based on descriptions.
Population Methods
Set values manually on individual accounts. Use AI enrichment to research and fill values automatically. Populate geographic data with location enrichment. Update values in bulk.
Monitoring & Refresh
Scheduled change detection flags when values shift. Ongoing monitoring keeps custom columns current. Automatic refresh for AI-enriched columns. Value normalisation and standardisation.
What Custom Columns Are
Custom columns are new data fields that you add to your territory account table. The key difference from standard enrichment: you define the question, and PG:AI answers it for every account. Want to know each account’s primary cloud provider? Their compliance posture? Whether they’ve been through a recent acquisition? You create a custom column, describe what you want to know, choose a data type (text, number, boolean), and PG:AI’s AI researches each account and fills in the answer. Custom columns can hold:- AI-researched answers - PG:AI looks at enrichment data, public information, tech stack, jobs, and more
- Manual values - data you set yourself on individual accounts
- Any data type - text, numbers, booleans - whatever fits the question
Viewing Custom Columns
In the Companies Table
Custom columns appear as additional columns alongside PG:AI Score, firmographic data, and enrichment columns. Each custom column shows the AI-generated (or manually set) value for every account.On Account Profiles
Click into any account to see all custom column values, giving you the full picture alongside standard enrichment data.Common Workflows
Getting a Specific Answer About Every Account
This is the core use case. You have a question that matters to your sales strategy, and you want the answer for every account in your territory.Create a custom column
Ask your admin to create a custom column (or create it yourself if you have access).
Set the data type
Text for open answers, number for quantitative data, boolean for yes/no questions.
| Question | Data type | Example output |
|---|---|---|
| What is their primary cloud provider? | Text | ”AWS”, “Azure”, “GCP” |
| Have they been through a recent acquisition? | Boolean | Yes / No |
| What compliance frameworks do they follow? | Text | ”SOC2, GDPR, HIPAA” |
| How mature are their DevOps practices? | Text | ”Early”, “Established”, “Advanced” |
| What is their estimated annual IT spend? | Number | 5000000 |
| Are they expanding internationally? | Boolean | Yes / No |
| What CRM do they use? | Text | ”Salesforce”, “HubSpot”, “Dynamics” |
| What’s their primary programming language? | Text | ”Python”, “Java”, ”.NET” |
The only limit is what you can describe clearly enough for AI to research.
Using Custom Column Answers for Prioritisation
Once columns are populated:Sort by custom column
Sort the territory by a custom column to see patterns - e.g., sort by “Primary Cloud Provider” to group all AWS accounts together.
Segmenting for Campaigns
Custom columns are powerful for campaign targeting:Filter for your target segment
Filter your territory to accounts at the right stage for your campaign message.
Adding Your Own Context Alongside AI Answers
Some columns work best as manual additions:- Account tier (Tier 1, 2, 3) - your team’s strategic classification
- Outreach status - “Not started”, “Sequenced”, “Meeting booked”
- Assigned rep - who owns this account
- Notes - context that only you have
Feeding Custom Columns into Scoring Models
Custom column values can be used as inputs to scoring models:- If “Primary Cloud Provider = AWS” and you sell an AWS-optimised product, that should boost the score.
- If “Recent Acquisition = Yes”, that might indicate budget disruption - positive or negative depending on your offering.
- If “DevOps Maturity = Advanced”, that could mean they’re ready for your product (or already have something similar).
How AI Answers Stay Current
Custom columns can be configured to refresh automatically:- Scheduled refresh - AI re-researches accounts periodically and updates values.
- Change detection - flags when an answer changes significantly (e.g., a company switches cloud providers).
- Manual refresh - re-run AI research on demand for specific accounts.
Tips and Best Practices
Choose the right data type - use boolean for yes/no questions, number for quantitative data, text for descriptive answers. The right type makes sorting and filtering work properly.
- Start with high-value questions - what data points, if you had them for every account, would most change how you prioritise? Start there.
- Combine AI columns with manual columns - AI tells you what’s true about the account; manual columns track your team’s decisions and status. Both are valuable.
- Review AI answers for accuracy - spot-check a few accounts you know well to validate that the AI is answering correctly. If not, refine the column description.
- Don’t over-create - 5–10 well-chosen custom columns are more useful than 30 random ones. Each column should earn its place on the table.
- Feed important columns into scoring - if a custom column captures a signal that genuinely affects fit, make sure it’s weighted in the scoring model.