Skip to main content
Agentic messaging gives each of your users their own AI lifecycle agent. Each agent reads that user’s in-app activity, decides whether a message would help, picks the right channel and moment, and writes the message itself, all inside your brand guidelines. There are no templates, journeys or sequences to build. You set the objectives and the rules, and agents do the work.
Agentic messaging is set up together with the Trophy team. Book a call to get started.

One agent per user, not one sequence for everyone

Traditional customer engagement platforms move users through segments and sequences: branching workflows of triggers, delays and templates that every user flows through. Every user on the same path gets the same message at the same step. Trophy treats each user as a cohort of one. Each agent works only for its user, so every message reflects that user’s own experience in your app. Take Mei, two weeks into a daily habit with a puzzle app. She’s still playing every day, but her sessions are getting shorter and she’s started playing late at night instead of on her commute.
  • A sequence-based platform waits until Mei has been inactive for three days, then sends a generic “We miss you, Mei!”.
  • A Trophy agent spots the early signs while Mei’s still active and gives her something new to work toward: “14 days in a row, Mei. You’ve been placed in your first weekly league with players at your level.”

How agents work

Agents pick the best message for each user at the best moment. Sometimes the best message is no message, and agents only reach out when they have something worth saying.

Grounded in product data

Agents see each user’s activity and progress in your app, from the metrics, events and attributes you send to Trophy.

Decide whether, when and where

Agents decide whether to send at all, pick each user’s preferred channel, and wait for the moment they’re most likely to act.

Every message written from scratch

No templates. Every message is written for one user, in your brand voice.

Get better every week

Agents learn from how each user responds, including opens, clicks and the ratings users give your emails.

What agents learn from

Agents use the same data that powers the rest of Trophy:
  • Metrics and events that describe how each user interacts with your app.
  • User attributes that you choose to share with agents in your brand kit.
  • In-app features like streaks, milestones and points, which give agents first-party signals to write about.
  • How each user engages with the messages they receive.
The more of your product you model in Trophy, the more context agents have to work with.

Agentic and legacy messaging

Agentic messaging replaces template-based emails and push notifications for the users it’s rolled out to. Each user receives either agentic messages or legacy messages, never both. Users in your control group receive no messages from Trophy, which lets you measure the impact of agentic messaging against a clean baseline.

Learn more

Objectives

Tell agents what success looks like for your product.

Brand kit

Teach agents about your app and how you sound.

Channels

Send email from your domain and push through your own providers.

Guardrails

The limits agents always work inside.

User preferences

Let users choose how much they hear from you.

Measuring impact

Prove uplift against a holdout group, not just clicks.

Rollout

How a rollout works, from integration to full launch.

Get support

Want to get in touch with the Trophy team? Reach out to us via email. We’re here to help!