Today, GTMOps (RevOps, BizOps, and MOps) teams at leading companies are not only incorporating Large Language Models (AI) into their workflow but also centralizing these transformations at the data warehouse and CRM-platforms level to supercharge their revenue outcomes. This blog shares the top 10 LLM prompts that help RevOps and MOps professionals gain deeper insights, personalize engagements, and streamline decision-making - and we’ve included more advanced GPT query examples to illustrate how each can be implemented using Census.
With customers spanning industries from B2B SaaS giants like Canva to B2C retail leaders like J.Crew and technical enterprises like Rippling, Census’s GPT Columns feature offers powerful new ways to execute these prompts directly on top of your customer data and sync the outcomes to tools like Salesforce, Hubspot, Braze and Marketo.
Note:
- The field names in following prompts are represented in Census’s Liquid Template format for ease of copy/paste
- Assumption: These prompts assume a generalized data model for leads, users, accounts and activity tables. Please, modify the prompts to fit your data model.
- You can model your customer data as Datasets in Census
Prompts in this guide:
1: Product-led Sales: Intelligent Lead Scoring for Sales Prioritization
Lead scoring can be as complex as we make. Here’s a simplistic version of a lead scoring prompt with just three parameters. Feel free to modify it to add more fields and additional business logic
Prompt:
“Analyze the lead’s industry, employee count, and website activity over the last 30 days. If they have visited the pricing or features page multiple times, increase the score by 10%. Assign a final score from 0 to 100 for prioritization.”
Fields
INDUSTRY: { { RECORD[‘INDUSTRY’] } }
EMPLOYEE COUNT: { { RECORD[‘EMPLOYEE_COUNT’] } }
WEBSITE ACTIVITY: { { RECORD[‘WEBSITE_ACTIVITY’] } }
Why It’s Valuable: For SaaS companies, lead prioritization is crucial. This prompt helps sales teams rank prospects based on engagement and readiness to purchase.
By centralizing this transformation in Census, teams ensure the scoring criteria are universally applied, providing consistency across all channels like Salesforce and Hubspot.
Popular Destinations: Salesforce, Hubspot, Attio, and Outreach
Assumption: Website activity is a json that contains, website visits with page urls, timeste stamps and overall summary of the website activity as low, medium and high.
2: Customer Success: Churn Risk Prediction for Retention Strategies
Prompt:
“Assess recent product usage, feedback scores, and support ticket history over the past six months. If recent activity has decreased by more than 50% or feedback is negative, classify the
customer as ‘High Risk’ and suggest follow-up actions.”
Fields
USAGE: { { RECORD[‘PRODUCT_USAGE’] } }
FEEDBACK SCORE: { { RECORD[‘FEEDBACK_SCORE’] } }
SUPPORT TICKETS: { { RECORD[‘SUPPORT_TICKETS’] } }
Why It’s Valuable: For RevOps teams, predicting and mitigating churn risk is crucial. This prompt supports targeted retention strategies based on customer engagement and feedback trends.
Popular Destinations: Gainsight, Churnzero, Planhat, Vitally, Hubspot, Salesforce
Assumption: Product Usage and support tickets are aggregated columns. Feedback score is closer to nps and other similar feedback scores. You can also use sentiment analysis here.
3: Hyper-Personalized Email Content for Upsells
Prompt:
“Analyze recent purchases, products browsed, and website activity in the last 30 days to create a personalized greeting and email recommendation. Suggest two new products based on their previous purchases and recent views, emphasizing complementary items.”
Fields
RECENT PURCHASES: { { RECORD[‘RECENT_PURCHASES’] } }
BROWSING HISTORY: { { RECORD[‘PRODUCTS_BROWSED’] } }
WEBSITE ACTIVITY: { { RECORD[‘WEBSITE_ACTIVITY’] } }
PRODUCT LIST: { { RECORD[‘PRODUCT_LIST’] } }
Why It’s Valuable: Personalized emails increase engagement and drive conversions. Census’s GPT Columns helps MOps teams deliver customized content at scale.
Popular Destinations: Braze, Iterable, Customer.io
Assumption: Data model is closer to a B2C business here. Website activity contains page visits
4: Industry Classification from Customer’s Website Data
Prompt:
“Use the website scraped and analyze its content to determine the industry it belongs to.
Focus on elements such as product offerings, services described, target audience, and terminology used throughout the site.
Based on your analysis, return the most appropriate industry label that best describes the primary focus of the website. Possible examples of industry labels include Technology, Healthcare, Finance, Education, Retail, Manufacturing, Real Estate, Hospitality, etc.”
Fields
WEBSITE_SCRAPED_DATA: { { record[WEBSITE_SCRAPED_DATA] } }
Why It’s Valuable: Correctly classifying companies enhances the accuracy of CRM data, supporting targeted outreach and better segmentation.
Popular Destinations: Braze, Hubspot, Salesforce, Outreach
Assumption: You already have scraped data from customers/prospects’ website or can get it.
5: Industry Categorization from Company Description
Prompt:
“Generate a short business industry name based on the following company description.
Possible examples of industry labels include Technology, Healthcare, Finance, Education, Retail, Manufacturing, Real Estate, Hospitality, etc.”
Fields
ORGANIZATION SHORT DESCRIPTION: { { RECORD[‘ORGANIZATION__SHORT_DESCRIPTION’] } }
Why It’s Valuable: Rapidly categorizing a company’s industry helps RevOps teams better understand market segments and target outreach accurately.
Popular Destinations: Braze, Hubspot, Salesforce, Outreach
Assumption: You already have customers/prospects’ description.
6: Customer Health Analysis
Prompt:
"Evaluate each account’s recent product interactions, support ticket history, and overall revenue contribution. Classify the account’s sentiment as ‘Excited,’ ‘Satisfied,’ ‘Neutral,’ or ‘Unhappy’ based on the following:
Excited: High product usage, minimal support tickets, and revenue contribution above average.
Satisfied: Moderate to high product usage, few support tickets, and average revenue contribution.
Neutral: Moderate product usage with occasional support tickets and below-average revenue contribution.
Unhappy: Low product usage, frequent support tickets, and low revenue contribution."
Fields:
PRODUCT USAGE: { { RECORD[‘PRODUCT_USAGE’] } }
SUPPORT TICKETS: { { RECORD[‘SUPPORT_TICKETS’] } }
REVENUE: { { RECORD[‘REVENUE’] } }
Why It’s Valuable: For Businesses to understand the state of their customers’ health and act before they become a churn risk.
Popular Destinations: Salesforce, Hubspot, Gainsight, Planhat, Customer.io, Braze, Iterable
7: Campaign Performance Analysis for Marketing Optimization
Prompt:
“Review engagement metrics (open rate, click-through rate) for recent campaigns targeting this customer. Come up with Engagement score for each campaign”
Fields:
OPEN RATE: { { RECORD[‘OPEN_RATE’] } }
CLICK-THROUGH RATE: { { RECORD[‘CLICK_THROUGH_RATE’] } }
Why It’s Valuable: Understanding campaign performance enables MOps to refine marketing efforts based on customer engagement insights.
Popular Destinations: Braze, Iterable, Customer.io
Note: The prompt can be modified to identify popular product and campaign messaging as well.
8: Behavioral Targeting
Prompt:
“Analyze purchase history, website visits frequency, and recent browsing behavior to classify the customer as ‘Frequent Shopper,’ ‘Discount Shopper,’ or ‘Luxury Shopper.’ Adjust segmentation based on recent sales participation or full-price purchases.”
Fields:
PURCHASE HISTORY: { { RECORD[‘PURCHASE_HISTORY’] } }
WEBSITE VISIT FREQUENCY: { { RECORD[‘WEBSITE_FREQUENCY’] } }
BROWSING BEHAVIOR: { { RECORD[‘BROWSING_BEHAVIOR’] } }
Why It’s Valuable: Targeted campaigns resonate more effectively. Census’s GPT Columns helps MOps teams segment audiences, delivering personalized messages aligned with shopper behavior.
Popular Destinations: Facebook Ads, Google Ads, Klaviyo
Note: Website frequency is recorded as low, medium, high, etc
9: Job Title Seniority Classification
Prompt:
“Classify the following job title into one of the levels of seniority: C-Suite, Executive Leadership, Senior Leadership, Management, or Individual Contributor. If the title is vague, default to Individual Contributor.
Fields
TITLE: { { RECORD[‘JOB_TITLE’] } }
Why It’s Valuable: Understanding a contact’s seniority allows MOps teams to target decision-makers effectively and tailor communications.
Popular Destinations: Outreach, Braze, Salesforce, Hubspot
10: Personalized Nurture Email based on Onboarding Progress
Prompt Example:
“Evaluate the setup steps completed and first-time product use within the onboarding period. Score onboarding progress as a percentage, highlighting any missed milestones, and provide suggestions for next steps. Draft a personalized email subject line and email content based on this.”
Fields
SETUP STEPS COMPLETED: { { RECORD[‘SETUP_STEPS_COMPLETED’] } }
FIRST-TIME PRODUCT USE: { { RECORD[‘FIRST_TIME_PRODUCT_USE’] } }
Why It’s Valuable: This prompt enables customer success teams to monitor onboarding, ensuring users reach key milestones promptly.
Popular Destinations: Braze, Iterable, Customer.io, Posthog, Amplitude, Mixpanel
Assumption: Setup Steps Completed and First Time Product Use are summarized column
By centralizing GPT-driven prompts through Census’s GPT Columns, RevOps, BizOps, and MOps teams can not only unlock insights at scale but also ensure those insights are consistent and reliably shared across all customer-facing applications. This approach makes AI-driven data transformation a single source of truth, streamlining data strategies to drive smarter decisions and deliver consistent, hyper-personalized experiences across every engagement channel.
Ready to elevate your data strategies? Explore GPT Columns in Census and start driving smarter data-driven decisions today.