What are infinity campaigns?

Summary

• Infinity campaigns are AI-driven, agentic marketing automation that continuously runs customer engagement strategies in real time without manual intervention or preset schedules
• While batch campaigns execute at fixed times, infinity campaigns evaluate each customer interaction as it occurs and take immediate action based on personalization rules, ML models, and business objectives
• Infinity campaigns adapt to customer behavior in real time, learning from engagement patterns and adjusting outreach, offers, and messaging on the fly
• Databricks CustomerLake enables infinity campaigns by combining a unified customer data platform with AI agents that orchestrate marketing actions, personalization rules, and reporting
• Use cases include dynamic offer personalization, real-time churn prevention, lifecycle automation, and continuous customer journey optimization

What are infinity campaigns?

Infinity campaigns are agentic marketing automation that runs continuously in real time rather than on a fixed schedule. Instead of defining a campaign that executes once or at preset intervals, infinity campaigns are always active, evaluating every customer interaction and deciding what action to take in milliseconds. The goal is to move marketing from discrete, periodic campaigns to a continuous feedback loop where each customer interaction is an opportunity to personalize, engage, or convert.
Traditional batch campaigns are scheduled: send email Tuesday at 10 AM to all customers matching a segment. Infinity campaigns are event-driven and persistent: when a customer visits your site, opens an email, or makes a purchase, the infinity campaign evaluates their profile and recommends the next best action immediately.

How infinity campaigns differ from batch campaigns

Aspect	Batch Campaigns	Infinity Campaigns
Execution	Scheduled (daily, weekly, on-demand)	Continuous, always active
Trigger	Time-based (e.g., every Tuesday)	Event-based (customer behavior)
Personalization	Static during campaign run	Updated in real time per customer
Adaptation	Rules set at campaign start	Rules adapt based on performance signals
Latency	Minutes to hours	Milliseconds
Scalability	Fixed audience size at start	Grows with new customer interactions

Batch campaigns are efficient for reaching broad segments at once. Infinity campaigns are better for 1-to-1 personalization and responding to immediate customer signals.

What makes infinity campaigns possible

Infinity campaigns require three technical capabilities:

1. Unified customer data (CustomerLake): A real-time, queryable view of each customer's profile, purchase history, preferences, and engagement signals, updated as interactions occur.
2. AI agents for orchestration (Agent Bricks): Autonomous decision-making systems that evaluate customer state, apply rules, and choose actions (send email, offer discount, pause contact, etc.).
3. Real-time personalization (ML models and rules engines): Trained models and business rules that score next-best actions and adapt based on customer response.

Without unified customer data, you can't know a customer's state across channels. Without agents, you can't make decisions automatically. Without personalization, you can't tailor action to each customer's context.

How Databricks approaches infinity campaigns

1. Build a unified customer data foundation

CustomerLake consolidates customer data from all sources (CRM, web analytics, transactions, support, email engagement) into a unified, queryable repository. CustomerLake runs on Databricks' lakehouse architecture, enabling both batch and real-time queries:

SELECT
  customer_id,
  current_segment,
  last_purchase_date,
  engagement_score,
  predicted_churn_risk,
  preferred_channel
FROM customer_profiles
WHERE customer_id = ?;

This query can run in milliseconds, returning the customer's current state to power immediate decisions.

2. Define rules and business logic

Infinity campaigns combine rules and AI:

• Rules define the boundaries: "Don't contact if they've unsubscribed," "Don't offer a discount if they purchased today."
• AI models score the next best action: which offer is most likely to convert, which channel is most effective, what timing maximizes engagement.

if customer_profile.engagement_score > 0.8:
    next_action = "send_personalized_offer"
    offer_type = ml_model.predict_best_offer(customer_profile)
elif customer_profile.churn_risk > 0.7:
    next_action = "send_retention_message"
else:
    next_action = "wait_for_signal"

3. Execute with Agent Bricks

Agent Bricks orchestrates the infinity campaign, executing decisions at scale:

• Agents continuously listen for customer signals (web events, purchase, email open).
• When a signal occurs, agents query the unified customer profile, apply rules, invoke ML models, and execute the action.
• Agents log all decisions for audit, performance tracking, and model retraining.

4. Monitor and optimize

Infinity campaigns generate continuous feedback: did the offer convert, did the message engage, did the retention action prevent churn. Use this data to:

• Measure performance (offer conversion rate, email open rate, retention rate).
• Retrain personalization models weekly or daily.
• Adjust rules based on what's working.
• Debug failures (why did this customer get this action?).

Use cases

• Dynamic offer personalization: Real-time pricing and offer selection based on customer behavior and inventory.
• Churn prevention: Detect at-risk customers immediately and execute retention actions (incentives, proactive outreach).
• Lifecycle automation: Automatically transition customers through lifecycle stages (onboard new, expand existing, retain at-risk, winback lapsed).
• Welcome and onboarding: New customers trigger personalized welcome journeys that adapt to their engagement.
• Continuous experimentation: Run A/B tests on offers, messaging, timing, and learn in real time.
• Cross-sell and upsell: Recommend products based on purchase history and real-time behavior.

FAQs

Are infinity campaigns available in general availability?

Infinity campaigns are currently in Private Preview on Databricks. Availability and pricing are subject to change; verify with your Databricks account team.

How do infinity campaigns differ from CDP campaigns?

Most CDPs execute batch campaigns on a schedule. Infinity campaigns are event-driven and always active. CDPs are reactive (activate a segment); infinity campaigns are proactive (respond to each interaction).

What data do I need to run infinity campaigns?

You need a unified customer profile with interaction history, purchase data, engagement signals, and any predictions or scores (churn, LTV, propensity). The more complete your customer data, the better the personalization.

How do I measure success of an infinity campaign?

Track key metrics for each action type: offer conversion rate, email engagement rate, push notification click rate, retention rate (for at-risk cohorts), and customer lifetime value impact. Compare treated customers to a holdout control group to measure true lift.

Can I run multiple infinity campaigns for the same customer?

Yes, but you need rules to prevent conflicting actions. For example, don't simultaneously send a discount offer and a retention message. Use priority or business logic to sequence competing actions.