Recommendation Systems & Personalisation

Every time a streaming service suggests a film, an online retailer recommends a product, or a news app curates your feed, you are interacting with AI recommendation systems. These are among the most commercially successful applications of AI, driving significant revenue for companies that can effectively match users with content, products, or services they are likely to want. The technology works by learning from your behaviour - what you click, buy, watch, and linger on - combined with patterns from millions of other users with similar preferences. For businesses, effective recommendations increase engagement, conversion rates, and customer loyalty. The technology is mature and its commercial value is well-proven. The concerns are about the effects on users and society. Recommendation systems optimised purely for engagement can create filter bubbles, amplify extreme content, and exploit psychological vulnerabilities. There is an inherent tension between recommending what users will engage with and recommending what is genuinely good for them. For your business, the choice of what to optimise your recommendations for - clicks, revenue, satisfaction, or some combination - has real consequences for your customer relationships.