When an AI platform that positions itself as a trusted source of information steps back from advertising, it is not a minor product tweak. It is a signal. Perplexity’s recent decision to deprioritise ads speaks to a growing unease around how AI’s monetisation strategies shape user perception – and ultimately, credibility.

For years, the dominant model of the internet has been simple: attract attention, monetise it through advertising, and optimise relentlessly for engagement. That model built some of the most valuable companies in the world – Facebook being a prime example. But generative AI introduces a different dynamic. These systems are not just entertainment feeds or retail marketplaces; they are positioned by their owners as trusted interpreters of the digital world, synthesizers of knowledge and, increasingly, decision-support tools in important business and life decisions.

Across the industry, there is growing recognition of the risk of what has been termed AI “enshittification” – the gradual degradation of user experience as monetisation pressures mount. Ads inserted into responses, prioritised commercial partnerships, or subtle ranking biases may generate short-term revenue, but they also create doubt. If an answer is useful, is it because it is accurate – or because it is sponsored?

For AI platforms that trade on authority, that ambiguity is dangerous. Trust, once compromised, is extraordinarily difficult to restore.

A Maturing Business Model for Generative AI

By prioritising subscriptions and enterprise revenues over advertising, Perplexity is signalling that long-term trust may be more valuable than short-term yield. This is not an ideological stance; it is a commercial calculation.

Subscription and enterprise models align incentives differently. When customers pay directly for a service, the provider’s revenue depends on continued satisfaction, reliability and performance. They expect clarity on how outputs are generated and what sources underpin them.

This represents a notable departure from the traditional ad-funded internet model. It suggests that generative AI may be entering a phase where differentiation is less about novelty and more about credibility.

There isn’t a single revenue model for AI. Consumer-focused tools may continue to experiment with ad-funded, freemium or hybrid approaches, as users often tolerate recommendations influenced by advertising. Enterprise users, however, have much lower tolerance for opaque monetisation: as AI becomes embedded in core decision-making processes, clarity and alignment with user interests is critical. Platforms that can demonstrate their commercial interests align with their users’ goals will have a structural advantage. 

This distinction matters across industries, particularly for organisations embedding AI into customer journeys or decision-support tools. Whether providing financial insights, product recommendations, or automated support, users expect transparency and unbiased outputs. If AI-driven insights appear influenced by commercial partnerships rather than objective data, trust in the platform can quickly erode. Subscription-based or outcome-aligned models create clearer accountability and reinforce the perception that a provider’s success is tied to its users’ success – a principle that has always been critical, and in the AI era, is becoming defining.