Co-pilot

In subscription-based businesses, knowing what to do for retention is hard. Knowing who to target, when, and how.,that’s where the real complexity lies.

How to setup the Co-Pilot

The Churned Co-Pilot turns that complexity into action.

Instead of relying on static rule-based flows or manual hypotheses, Co-Pilot uses a reinforcement learning engine to experiment, learn, and optimize the way you engage your customers, automatically.

Workflow builder Churned example. Based on Health change and using to push a Klavyio flow

From Exploration to Exploitation: A Smarter Way to Run Your Retention Playbook

Co-Pilot operates in two dynamic phases:

1. Exploration Phase: Learning What Works

During this phase, Co-Pilot randomly distributes multiple treatments (like discounts, win-back emails, or in-app nudges) across user segments to measure impact. Think of it as automated A/B (or even A/B/n) testing, but without requiring constant input from your team.

For each customer journey stage—onboarding, renewal, post-churn—Co-Pilot tracks how different interventions influence key outcomes: retention, reactivation, upsell, and engagement.

Example: Instead of guessing whether a “15% discount” or a “personalized win-back email” works better, Co-Pilot tests both and collects conversion data.

2. Exploitation Phase: Acting on the Best Strategy

Once enough signal is gathered, Co-Pilot shifts gears. It starts applying the best-performing action to each customer segment, based on historical response data and predicted uplift. This is where exploration turns into ROI.

Your workflows go from generic to precision-targeted—on autopilot.

For one user, the best next step might be a WhatsApp message with a survey. For another, it’s a discount via email. Co-Pilot figures it out—and does it.

🔄 How It Integrates With Your Customer Workflows

Co-Pilot doesn’t replace your journey builder. it powers it.

• You define the allowed actions per flow (e.g., message variants, incentives, channels).

• Co-Pilot plugs into these steps as the decision engine.

• It tracks every outcome, adjusts decisions in real time, and pushes the right action at the right moment via your marketing tools (e.g., HubSpot, Customer.io, Klaviyo).

It’s compatible with onboarding flows, win-back campaigns, engagement loops, and renewal sequences, especially when you're working with large user bases and limited human bandwidth.

📈 Why It’s Better Than Just A/B Testing

Traditional A/B testing is static. You test one thing, wait weeks, pick a winner, and repeat.

Co-Pilot is adaptive. It runs multi-arm bandit tests, continuously learns from new data, and shifts budget or attention to the best actions in real time. It scales better than any manual setup.

No more campaign fatigue. No more one-size-fits-all. Just evolving, intelligent workflows.

⚙️ Use Cases Where Co-Pilot Excels

• Churn Prevention: Test incentives, content types, and timing. Automatically apply what works best to each churn risk segment.

• Reactivation: Learn which nudges convert dormant users back into paying ones.

• Engagement Boosting: Identify the most effective engagement hooks across behavioural cohorts.