← All solutionsSolution · Retention and monetization

Keep the users your AI feature brings in

AI apps earn more per user and churn faster. We work on the moments where users decide to stay: where the offer sits, what the trial ends on, what leaving looks like, and how credits and plans are priced against what inference really costs.

Who it is for

Live consumer apps whose AI feature gets tried once and dropped, or whose subscribers leave after the first renewal.

What we can build for you

Six things we can put in your product, each backed by work we have already shipped.

01
Offers placed after value
Paywalls and trials shown once a user has used what they unlock, never as a wall in front of it.
02
Churn measured properly
By signup cohort, with AI users compared against everyone else.
03
Experiments with real statistics
Sample sizes, significance tests and split checks, shipped through Remote Config.
04
Credits and hybrid pricing
Packs and plans priced against real inference cost, on RevenueCat.
05
Store or web checkout
What goes to Apple and Google, and when a web checkout earns you more.
06
Attribution you can trust
Purchases reported from the server, matched to the ad click, never counted twice.

The four tests, for this

Where this kind of feature usually breaks, and what passing each test looks like.

01 · Design
Does it feel right?

Offers read as part of the product, and leaving is honest, not a maze.

02 · Mobile and backend
Does it hold up?

Purchases, trials and entitlements stay correct across devices, stores and the web.

03 · AI, retrieval and evals
Is it right?

Every change judged on renewals by cohort, not on the day's conversion rate.

04 · Metering and pricing
Most at risk
Does it pay?

Prices set against what each user really costs you to serve.

Closest track: Stop the churn

How the two weeks go

€6,000, fixed. Written findings on day five; stop there and pay half. You keep everything either way.

  1. Days 1 to 4Measure it: retention of AI users against everyone else, by cohort, and where in the flow they leave.
  2. Day 5Findings: which moment costs you the most users, and what fixing it is worth.
  3. Days 6 to 9Fix that moment behind an experiment, so the result is measured rather than argued.
  4. Day 10Handover with the experiment live and its readout date set.

Where we have done it

Previous work · 2024 to 2025
AI learning app at 30M users
Monetization features that helped cut churn by 35% on an app with 30M users.
Built by our team · 2025 to now
Voice AI writing app
Trials, a yearly offer on paywall close, and purchases attributed from the server.
Built by our team · 2026, early access
AI product analytics platform
Experiments with sample sizes and split checks, watched after shipping.

Tell us what you are building.

Thirty minutes, no deck. You get a straight answer about whether we are the right people, and we say no when we are not.