Jun 16, 2025
Schaun Wheeler

How Aampe's Agents Use Causal Analysis to Measure Impact Amidst External Messaging

Jun 16, 2025
Schaun Wheeler

How Aampe's Agents Use Causal Analysis to Measure Impact Amidst External Messaging

Jun 16, 2025
Schaun Wheeler

How Aampe's Agents Use Causal Analysis to Measure Impact Amidst External Messaging

Jun 16, 2025
Schaun Wheeler

How Aampe's Agents Use Causal Analysis to Measure Impact Amidst External Messaging

One of our customers asked a great question:

If a user is getting messages from systems outside Aampe (this can happen for many different reasons), how can we tell which outcomes were caused by that user's agent? Wouldn’t those other messages interfere with the agent’s learning?

This question points to an important aspect of how our agents work: they don’t rely on simple associations. Rather, they do causal analysis.

When an agent sends a message, it doesn’t just look at what the user did afterwards. It also looks at:

  • What that user had been doing in the hours before the message

  • What’s normal for that user overall

  • What’s typical behavior across the entire app population

Only after subtracting out those baselines does it take credit for any observed change. That’s how the agents isolate uplift and make good decisions going forward.

But what about other messages — like transactional or support outreach?

There are two scenarios:


  1. If those messages are closely tied to in-app events (e.g., a transactional message triggered by a purchase), that signal is already embedded in the event stream. So the agent naturally incorporates that context into its baseline, even without knowing there was a corresponding message sent.

  2. If those messages aren’t tied to observable events, then we can explicitly add them to the event stream so that agents can reason about them. The system is designed to ingest and respond to any observable factor, not just the ones it initiates.

In other words: the agent is always trying to understand what its own actions changed, in a noisy world full of other influences. That’s what makes it agentic.

Personally, I think it's a good idea for all messages — transactional or otherwise — to be orchestrated agentically. But Aampe agents can learn robustly even when they don’t control the full message share.

Because the agents aren't really delivering content - that's just the mechanism for interacting with the user. Those interactions form the basis for reasoning about impact. That's what the agents are really doing.

0

Related

Shaping the future of marketing with Aampe through innovation, data.

Renewals, holidays, and launches don’t need hardcoded rules. With reward signals, eligibility criteria, and timing action sets, agents adapt naturally to recurring patterns.

Renewals, holidays, and launches don’t need hardcoded rules. With reward signals, eligibility criteria, and timing action sets, agents adapt naturally to recurring patterns.

Renewals, holidays, and launches don’t need hardcoded rules. With reward signals, eligibility criteria, and timing action sets, agents adapt naturally to recurring patterns.

Renewals, holidays, and launches don’t need hardcoded rules. With reward signals, eligibility criteria, and timing action sets, agents adapt naturally to recurring patterns.

Aug 21, 2025

Schaun Wheeler

By modeling statistical relationships between events, agents evaluate directional shifts in behavior—so the same system adapts across every lifecycle stage.

Aug 21, 2025

Schaun Wheeler

By modeling statistical relationships between events, agents evaluate directional shifts in behavior—so the same system adapts across every lifecycle stage.

Aug 21, 2025

Schaun Wheeler

By modeling statistical relationships between events, agents evaluate directional shifts in behavior—so the same system adapts across every lifecycle stage.

Aug 21, 2025

Schaun Wheeler

By modeling statistical relationships between events, agents evaluate directional shifts in behavior—so the same system adapts across every lifecycle stage.

Aug 19, 2025

Schaun Wheeler

You don’t coach by chasing the trophy. You coach by tracking whether each play puts you in a stronger position. The same is true for customer engagement.

Aug 19, 2025

Schaun Wheeler

You don’t coach by chasing the trophy. You coach by tracking whether each play puts you in a stronger position. The same is true for customer engagement.

Aug 19, 2025

Schaun Wheeler

You don’t coach by chasing the trophy. You coach by tracking whether each play puts you in a stronger position. The same is true for customer engagement.

Aug 19, 2025

Schaun Wheeler

You don’t coach by chasing the trophy. You coach by tracking whether each play puts you in a stronger position. The same is true for customer engagement.

Jul 23, 2025

Schaun Wheeler

A/B tests help us see what works on average, but real users aren’t average, their motivations and contexts vary. That’s where agentic learning shines, adapting to individuals over time. The best results come when we layer the two: tests for clarity, agents for personalization.

Jul 23, 2025

Schaun Wheeler

A/B tests help us see what works on average, but real users aren’t average, their motivations and contexts vary. That’s where agentic learning shines, adapting to individuals over time. The best results come when we layer the two: tests for clarity, agents for personalization.

Jul 23, 2025

Schaun Wheeler

A/B tests help us see what works on average, but real users aren’t average, their motivations and contexts vary. That’s where agentic learning shines, adapting to individuals over time. The best results come when we layer the two: tests for clarity, agents for personalization.

Jul 23, 2025

Schaun Wheeler

A/B tests help us see what works on average, but real users aren’t average, their motivations and contexts vary. That’s where agentic learning shines, adapting to individuals over time. The best results come when we layer the two: tests for clarity, agents for personalization.

Load More

Load More

Load More

Load More