As artificial intelligence becomes increasingly embedded in digital infrastructure, platform-based businesses face both unprecedented opportunities and strategic challenges. From enhancing network effects and personalizing user experiences to reshaping competitive dynamics and governance models, AI is no longer a peripheral tool, it is becoming central to how platforms grow, differentiate, and sustain their market positions. In this Q&A feature, Andrei Hagiu, Professor of Information Systems explores how AI intersects with core platform concepts such as feedback loops, multihoming, trust, and disintermediation, offering insight into how platform firms can adapt and thrive in an AI-driven landscape. Belfort Group’s Drew Vaughan poses the following questions:
Your research highlights the importance of network effects in platforms. How can AI amplify these effects to drive growth?
AI can enhance both “traditional” and data-driven network effects by improving matching and transaction efficiency on platforms. As platforms gather more user data, AI models can deliver increasingly relevant recommendations, search results, and matches, thus making the platform more valuable to users (buyers and sellers if we are talking about marketplaces). These improvements can attract even more users, fueling a feedback loop of engagement and learning. The key is to figure out what feedback signals can be gleaned from users (e.g., how many other users they engage with and for how long, failed transactions, etc.), which can be fed into the AI models to help improve the experience.
How can AI assist platforms in maintaining quality and trust as they scale?
AI can help platforms maintain quality and trust at scale by automating moderation and detecting fraud or misinformation in real time. For instance, recommendation systems and generative models can be fine-tuned to flag harmful content or guide users toward higher-quality interactions. However, it is important to realize that the effectiveness of AI in doing so depends heavily on the quality of data feedback loops, which in turn depend on the ability to extract reliable signals from user interactions. If signals are unreliable, that limits AI’s effectiveness and means the platform still needs to rely mainly on careful mechanism design and human intervention.
Looking ahead, how do you envision AI reshaping the competitive landscape among platform businesses in the next five years?
AI is likely to erode incumbent advantages and enable new forms of service delivery. AI agents could replace or bypass traditional platform interfaces—such as Booking.com or Uber—by directly performing tasks for users, thereby weakening network effects and reducing “multihoming” costs (i.e., difficulties in joining and using multiple platforms). At the same time, platforms that integrate AI deeply and personalize services through strong feedback loops may sustain their dominance. The result could be a bifurcation: some incumbents disrupted by AI-native challengers, and others evolving into new forms of gatekeepers through AI integration.
Is AI turning platforms into commodities—or is it giving them tools to offer more tailored, differentiated experiences?
Overall, I think AI gives platforms new ways to differentiate—through personalized and adaptive user experiences. Think about what Airbnb will look like once it transitions to a fully interactive (gen AI-based) interface. Again, the key is to embed proprietary data and learning feedback loops that increase switching costs and user lock-in. Platforms that can harness both across-user and within-user learning are particularly well-positioned to offer distinctive value. That being said, AI agents may become the new dominant platforms, relegating some of the existing big platforms to a background role.
What are the most effective strategies existing platforms can use to embrace AI without cannibalizing their business models?
First, existing platforms should embed AI tools (e.g., copilots, chatbots, personalized assistants) into existing services to increase engagement, efficiency, and user retention. That can be done without undermining the platform’s core revenue sources (advertising or commissions). For example, Google has cautiously incorporated AI Overviews into search results while maintaining ad visibility. Second, platforms can design their own AI systems to leverage unique proprietary data and reinforce their data feedback loops. This should lead to higher switching costs and a lower likelihood of disintermediation. Done right, AI can deepen user relationships and expand a platform’s value proposition without replacing its underlying business model. There are of course exceptions. For instance, it is hard to see how Uber can manage the transition to autonomous vehicles without significantly changing its existing business model that relies on independent human drivers.
