How Popsa used Amazon Nova to inspire customers with personalised title suggestions

Popsa, a technology company that transforms everyday photos into printed Photo Books, has leveraged Amazon Bedrock and the Amazon Nova family of models to reimagine its Title Suggestion feature. Originally launched in 2021, the feature used metadata and on-device convolutional neural networks to generate simple titles like "France 2024." By integrating Amazon Nova Lite, Amazon Nova Pro, and Anthropic's Claude 3 Haiku via Bedrock's unified API, Popsa now automatically generates creative, brand-aligned titles and subtitles across 12 languages. The system combines metadata (timestamps, geocoordinates), computer vision (detecting beaches, barbecues, pets), and retrieval-augmented generative AI to produce outputs that meet strict requirements: titles and subtitles under 36 characters, valid JSON format, and appropriate category icons. This resulted in over 5.5 million personalised titles generated in 2025, improved quality, reduced costs, and faster response times, leading to higher customer satisfaction and measurable uplifts in engagement and purchase rates.

Popsa's core philosophy is that technology should do the heavy lifting for users, and this update exemplifies that approach. The company built an evaluation pipeline with over 100 example Photo Books to ensure outputs met criteria like character limits, theme consistency (e.g., no skiing icons for beach holidays), brand style, and title-subtitle cohesion. The generative AI solution replaced a rule-based algorithm called Title Suggestion Graph, which offered limited creativity. Now, users receive suggestions that are more inspirational and personalized, elevating the design of their Photo Books. Available across 50+ countries, Popsa's use of Amazon Nova showcases how smaller tech companies can apply cutting-edge AI to enhance user experiences without complicating the software. The results demonstrate the power of combining metadata, computer vision, and generative AI to solve real-world problems in consumer products.