Industry
Retail
Omnichannel CX and Intelligent Innovation – The Future of Retail
Highlights
- Generative AI’s (GenAI) out-of-the-box capabilities and its ability to generate human-like content and responses represent a paradigm shift in artificial intelligence (AI).
- From reimagining operations to democratizing expert knowledge and accelerating innovation, GenAI can create exponential value across the retail enterprise.
- The synergy between AI and GenAI presents an exciting frontier in retail. By combining both their strengths, retailers can build intelligent enterprises of the future.
On this page
Looking under the hood of GenAI
By putting artificial intelligence (AI) in the hands of common people, generative AI (GenAI) has stoked tremendous interest and upended the way they interact with the technology.
According to our recent report, the TCS AI for Business Study, 47% of retail executives expect up to half their employees to be using GenAI on a daily basis within the next three years. To leverage this new technology alongside their existing AI investments, it is important for retailers to look under the hood and understand the practical applications it has today and what the future holds.
Broad applications in an enterprise context
Undoubtedly, GenAI is highly disruptive, very nebulous, and has the potential to be greatly transformative.
Large language models (LLMs) and large action models (LAMs) represent two powerful categories of GenAI models that have a wide range of applications for retailers. They can help optimize operations, innovate, and elevate the customer experience. We explore a few GenAI retail use cases and how GenAI can work alongside traditional AI to generate value for businesses.
Reimagining operations to accelerate time to market
LLMs, with their potential multi-modal capabilities, are adept at processing natural language, decoding consumer sentiment, summarizing content, and generating autonomous content. They can revolutionize retail operations. For example, they can automate and transform digital commerce operations such as item attribute extraction and product copy creation and validation, thereby reducing product onboarding time by up to 80% and lowering operational costs by up to 30%. By reducing the time taken to analyze customer sentiment from weeks to hours, GenAI has sped up the process of leveraging the voice of the customer to develop targeted sales strategies, enhance products, and introduce new product lines, improving net promoter score (NPS) by 15–20%. E-commerce sites can speed up product discovery by leveraging chatbots to comb through frequently asked questions (FAQs) and terms and conditions (T&Cs) sections of websites, giving shoppers quick access to information such as returns and exchange policies or sizing in the case of lingerie and shapewear. In stores, GenAI-powered product guides can help store associates close more sales and increase average order value by equipping them with comprehensive product knowledge.
LAMs, an advancement over LLMs, can act based on natural language instructions and automate workflows while keeping humans in the loop before taking critical actions. For example, LAMs can review customer shoutouts, classify them, and trigger resolution workflows based on the severity level.
Creating exponential value with democratized expert knowledge
A significant amount of tacit domain knowledge can potentially be digitized into LLMs as they are finetuned for enterprise use. This democratizes expert knowledge and enables decision-makers to get on-demand access to it. Additionally, GenAI-powered virtual assistants can free decision-makers from the constraints of traditional dashboards by giving them more power to interact in natural language with the output of enterprise AI models and get recommendations along with visualizations. For example, merchandising teams can get a high-quality output by entering prompts such as:
- Give me pricing recommendations for at least a 4% uplift in sales for the cereals category.
- Can you recommend a promotional campaign to increase footfall?
LLMs can tap into unstructured data such as store’s standard operating procedures (SOPs), training manuals, and employee policies to enable faster associate onboarding. They can also analyze supplier contracts to identify key terms such as pricing, delivery schedules, and payment terms, making it easier for retailers to monitor compliance and negotiate better terms.
Accelerating innovation by augmenting human creativity
Identifying emerging trends with a global resonance while overcoming creative plateaus is a continual challenge for fashion retailers. With GenAI, they can analyze vast datasets encompassing past collections, social trends, fashion blogs, street style from key fashion capitals, and upcoming celebrity events to autonomously curate mood boards and generate novel design concepts.
From a customer acquisition perspective, GenAI in retail can help digital marketers generate deeply persuasive and creative content for microsegments by evaluating trending topics, generating blogs and articles, embedding high-ranking keywords into site content, and localizing content and images. It can also help them bid for the right keywords, understand email click-through rates, and run campaigns effectively. Better search visibility can increase click-through rates by 2% and the average basket size by 20%.
GenAI in retail: Spectrum of current and future potential
The out-of-the-box capabilities of GenAI models are quite vast in areas such as content generation, conversational interfaces, and knowledge discovery.
At this point in time, GenAI cannot solve optimization and estimation problems or test statistical hypotheses by itself. However, when combined with other AI techniques, it can add value by offering better accuracy, transparency, and performance, while also reducing costs and the need for data. Examples include automating the components or steps involved in solving demand forecasting, supply planning, and transportation optimization problems such as data cleansing and transformation, feature selection and engineering, and generation of synthetic data.
GenAI can also attempt to depict underlying real-world scenarios through problem definition and assist in mathematical formulations. But it cannot reason coherently or come to a decisive conclusion under all contexts, and is prone to hallucinations. Having domain experts in the loop will help LLMs perform better over time. As the technology matures, the degree of human intervention is likely to reduce, expanding potential use cases to core retail areas such as merchandising and supply chain operations.
Adopting a synergistic approach
The synergy between AI and GenAI presents an exciting frontier for retail.
Although the cost of training GenAI models in-house today is exceptionally high, with time, LLMs are likely to be ubiquitously available as open-source models, paving the way for wider adoption as virtual agents that interact with humans in retail workflows. By combining the strengths of AI and GenAI, retailers can build intelligent retail enterprises of the future. The top three concerns for GenAI adoption are the need for an ecosystem of technology, data, and model partners to deploy it cohesively; the implications of a shift toward automation and efficiency; and security and privacy. It is crucial to balance the benefits of GenAI with ethical concerns by adopting a strategic approach that prioritizes responsible AI development and deployment.
