With users increasingly favouring AI assistants for routine tasks over traditional apps, Gartner forecasts a 25% decline in mobile app usage by 2027, signalling a transformative shift in how software is accessed and utilised.
The case for AI assistants replacing some traditional apps is no longer theoretical. Users are increasingly moving from tapping through separate services to issuing a single instruction and letting software coordinate the rest. According to a Gartner forecast published in January 2025, mobile app usage could fall by 25% by 2027 as assistants from Apple, Google, Meta and OpenAI take over more routine tasks. Gartner also estimated that agentic AI could place as much as US$234bn in application software spending at risk by 2030.
That shift is already visible in consumer behaviour. TELUS Digital said its survey found that 32% of consumers had replaced at least one app with an AI assistant over the past year. The company said respondents were attracted by speed and personalisation, with 83% saying AI completed tasks faster and 71% saying it adapted better to their preferences over time. That suggests the appeal is not only convenience, but also a sense that conversational systems are becoming more useful than fixed app interfaces for everyday work.
The likely outcome is not the end of software applications, but a change in where users encounter them. Many products may continue to operate in the background, with their databases, APIs and workflows still doing the heavy lifting while the assistant becomes the main point of entry. This is consistent with the broader move towards agentic systems that can interpret intent, plan a sequence of actions and call external tools. Google and other platform developers are pushing in that direction by embedding more AI-driven automation into the operating system itself.
Even so, app interfaces still have clear advantages in areas that depend on precise control, dense information and visual interaction. Video editing, gaming, social media feeds and advanced analytics are all examples where a conversation is unlikely to replace a full screen of tools. Trust remains another constraint. Granting an assistant permission to buy, edit or transfer data introduces obvious security and privacy risks, and Gartner has warned that many current agentic AI projects are still not mature enough for complex tasks. The more realistic near-term future is therefore one of fewer app launches, not no apps at all. Users may interact with a thinner layer of software while the underlying services remain intact.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
