Report finds AI apps face challenges keeping users engaged over time

With app stores now crowded with AI apps, many developers may assume that the clearest path to making money is to add artificial intelligence features to their own products. But a new report focused on the subscription app economy across iOS, Android, and the web suggests that assumption may not hold up as neatly as expected.

RevenueCat, a company whose subscription management tools are used by more than 75,000 app developers, said in its 2026 State of Subscription Apps Report that integrating AI does not automatically translate into long-term subscriber loyalty. Instead, the report found that AI-powered apps have a harder time retaining paying users over time, with people cancelling annual subscriptions — a metric known as churn — 30% faster than users of non-AI apps at the median.

The report is based on an analysis of subscription app businesses that rely on RevenueCat’s tools to process more than 1 billion in in-app transactions, generating over $11 billion in developer revenue each year. Because RevenueCat is one of the more widely used tools in this category, the company says its data offers a strong sample for identifying broader trends.

Among the more notable findings, the report said that the majority of apps using RevenueCat’s platform are still not AI-powered. Across all categories, AI-powered apps make up 27.1% of apps, while non-AI apps account for the remaining 72.9%. Even so, AI is clearly becoming more common, with roughly one in four apps now marketingthemselvesf as AI-powered.

For clarity, the AI-powered category includes not only major AI chatbot apps such as ChatGPT and Gemini, but also any app that describes itself as being powered by AI.

Photo and Video is the category with the largest share of AI-powered apps, at 61.4%, while gaming has the smallest presence, at just 6.2%. Travel, at 12.3%, and Business, at 19.1%, also remain relatively low-AI categories.

The more striking findings, however, concern how well AI apps hold onto paying users. According to RevenueCat’s data, AI apps perform worse on retention at both monthly and yearly intervals.

Annual retention — the share of subscribers still paying after 12 months — stood at 21.1% for AI apps, compared with 30.7% for non-AI apps. Every month, AI apps showed retention of 6.1%, versus 9.5% for non-AI apps, a gap of 3.4 percentage points.

The only timeframe in which AI apps outperformed was the weekly subscription window, where they posted retention of 2.5% compared with 1.7% for non-AI apps. Even so, weekly subscriptions are not the most common option among AI-powered apps.

These trends may reflect how quickly AI technology is changing. Users may be moving between AI apps more rapidly than in other categories, testing different products in search of the one with the latest, most capable underlying models.

As people try more AI apps, they are also more likely to find that some of them fail to deliver enough value. The report says AI apps have refund rates 20% higher than non-AI apps, with a median refund rate of 4.2% versus 3.5%.

The upper limit of refund rates is also higher for AI apps, reaching 15.6% compared with 12.5% for non-AI apps. According to the report, which points to “greater volatility in realised revenue and deeper issues in user value, experience, and long-term quality.”

Still, the datasuggeste clear advantages to being in the AI app category.

RevenueCat found that AI apps convert users from trial to paid subscriptions 52% better than non-AI apps, with a median conversion rate of 8.5% compared with 5.6%. AI apps also monetise their downloads around 20% more effectively than non-AI apps, with a median of 2.4% versus 2%.

AI apps also deliver 39% higher monthly realised lifetime value (RLTV), a metric that measures the actual net value of an average paying user over time. Every month, the median RLTV for AI apps was $18.92, compared with $13.59 for non-AI apps. On an annual basis, AI apps posted RLTV that was 41% higher, with a median of $30.16 versus $21.37.

The broad conclusion from the report is that AI can be highly effective at driving early monetisation. However, apps built around it are still struggling to sustain their value proposition for users over the long term.