As artificial intelligence (AI) continues to advance, organizations are looking for ways to unlock its value and move from proof of concept (POC) to real-world impact. A recent article by APAC Chief Technology Officer for Thoughtworks on GovInsider discusses the challenges and strategies involved in this process [7d504c44]. The article highlights that as many as 90% of AI and generative AI (GenAI) projects are stuck in the POC stage and not productionized. Thoughtworks encourages organizations to start testing AI with use cases emerging in their operations. High-quality labeled data and data access are common challenges preventing companies from deploying AI models in production. Thoughtworks provides tools and expertise to evaluate Large Language Models (LLMs) and offers accelerators for tasks like text classification and data labeling. The article also provides examples of successful AI implementations, such as the yuu Rewards Club in Singapore and a leading South Asian bank. These organizations were able to swiftly progress from POC to full-scale production by having strong leadership endorsement and dynamic GenAI strategies. The article emphasizes the importance of establishing a responsible AI framework that addresses privacy, security, and compliance with laws and regulations. Overall, the article offers valuable insights and recommendations for organizations looking to unlock the value of AI and achieve real-world impact.
In the race to develop cutting-edge AI experiences, enterprises are pouring resources into various models and technologies. According to experts from Pinterest, LinkedIn, and Slack, the key to creating an AI product that truly meets customer needs is cross-functional collaboration. Deepak Agarwal, VP of engineering at Pinterest, emphasized the importance of approaching AI product development with an AI-first mindset and establishing a culture of collaboration among engineering, design, product, data, and legal teams. The development of AI products presents challenges due to the non-deterministic nature of generative AI, which requires developers to focus on innovation, quality, safety, and performance. Rapid prototyping and iterative development have become essential in this environment. However, companies often fail to bring together the teams working on AI and those responsible for assessing risks and compliance issues. To bridge these gaps, experts recommend collaborating across functions, prioritizing customer needs, and triangulating feedback from various sources before jumping into development and deployment. [7ff344a8]
Product taxonomy plays a key role in AI-driven retail strategies. To harness the power of AI for better context, natural language search, predictive analytics, and hyper-personalization, retailers need a strong taxonomy structure for their product catalogs. Product taxonomy organizes products in a hierarchy of attributes and relationships, creating a common language between humans and machines. A comprehensive product taxonomy includes multiple levels of classification, ensuring accurate categorization for various retail processes. Robust product taxonomy offers benefits such as improved search accuracy and enhanced navigation. When integrated with AI, product taxonomy enhances customer experience by enabling better site search, product discoverability, and hyper-personalization. Building a system of product taxonomy requires understanding the product mix and customer perspectives, establishing clear data governance policies, designing the taxonomy with the customer experience in mind, and continuously refining the system based on feedback. A well-designed product taxonomy harmonizes with AI to create the best product experience for customers. [9f7c4b41]