Generative AI: The future of e-commerce is here

Artificial intelligence (AI) makes online shopping a hyper-personalised experience for consumers, while on the backend, it streamlines supply chains and optimises inventory. Online retailers, top brands, and specialised vendors are quickly evolving to embrace AI, opening the door to investment opportunities. Today’s leaders will be best positioned for tomorrow. Those leading the way today will be best positioned for tomorrow, with U.S. e-commerce revenue alone expected to grow 53% between 2023 and 2027, reaching $1.4 trillion USD.¹  Globally, online sales are expected to make up a quarter of total retail by 2027.²  Yet the costs of shipping, returns, price matching, and supply chain disruptions have led to shrinking profit margins. AI could boost sales while maximising logistical efficiency — giving early adopters the potential for a competitive edge.

On the front end, AI serves as a “sales assistant” for every consumer, from product discovery to fielding questions. Behind the scenes, AI can analyse data in real-time to optimise logistics and improve supply chain efficiency. We believe this disruptive growth potential should cause investors to note the top tech companies, specialised start-ups, and leading retailers innovating with AI today. 

What is generative AI for e-commerce?

Generative AI is a specific type of AI, used to create something new based on a given set of training data. Its growth story is driven by a combination of capabilities, and the tech enablers supporting these. One capability is machine learning (ML). This method of data analysis uses algorithms to parse data, learn from it, and make predictions or decisions with minimal human intervention. Generative AI is a specialised version of ML, focused on generating new instances, like text or images. Computer vision involves teaching machines to interpret and make decisions based on visual data. It powers visual search, product recognition, and indexing and tagging processes. Image generation is a capability which allows retailers to quickly generate high-quality, relevant imagery to showcase products and appeal to specific demographics. Platforms like Dall-E from OpenAI and Midjourney are current leaders in image generation. Natural language processing (NLP) refers to how programs can understand and generate language, which powers chatbots, search functions, sentiment analysis, and product recommendations. Hugging Face, a company which started in NLP, is a multi-billion company which has attracted investments from Nvidia, Salesforce, and others.³

On the tech enablers side, datasets are collections of information, such as data on customer behaviour, sales trends, or product information, used to train and validate AI or ML models. Large language models (LLMs) are trained on vast amounts of text data to generate human-like text, enabling various content creation capabilities. Arguably the most well-known LLM is GPT, used in ChatGPT, and trained on 1.8 trillion parameters.⁴   Graphics processing units (GPUs) comprise of a semiconductor chip designed to handle complex calculations and parallel processing tasks. Nvidia proved to be a leading GPU provider in 2023, while Google has launched cloud tensor processing units (TPUs) as an alternative.⁵

Personalised consumer journeys – the new user experience

Generative AI analyses customer data and behaviour patterns to enhance and personalise the consumer experience, driving online sales and retention. Targeted marketing and advertising allows retailers to use AI tools and consumer data to target customers based on behaviour and demographics, offering tailored marketing campaigns and discount offers pinpointed to encourage engagement and loyalty. Meta has unveiled its AI sandbox with AI-driven ad updates for advertisers. Rival Google announced a new Search Generative Experience feature built to perform smart shopping research for users and enable seamless purchases. Startup Treat uses generative AI with proprietary and customer data to create personalised ads that incorporate elements shown to perform well with target demographics. One study shows that 66% of consumers are willing to share their data in exchange for cheaper options.⁶

Today’s conversational chatbots are next level. With NLP capabilities, they are context-aware sales assistants, tailoring product discovery, recommendations, and deals in real time. They can cross-sell and upsell, assist with logistical enquiries, and gather feedback. Some retailers have launched their own chatbots, such as online resaler Mercari’s Merchat AI, powered by ChatGPT. Carrefour was one of the first to integrate OpenAI into its customer experience, with a chatbot that helps shoppers with product selection. Other retailers rely on vendors, such as established players in omnichannel retail like IBM’s Watson Assistant and Salesforce’s Contact Genie, or more specialised companies such as Inventa Holdings, Inc., eGain Corporation, Nuance Communications, and Conversia, Inc., that focus on purpose-built bots. While ChatGPT is popular, other NLP solutions could soon appear in e-commerce. Inflection AI, which offers AI-powered chatbot Pi, is a multi-billion dollar company funded by Microsoft and Nvidia.⁷  Competitors such as Character.AI are also emerging.⁸

Product design tools leveraging generative AI can spark consumer creativity, allowing them to customise products for direct purchase, or help designers rapidly develop new prototypes and stay ahead of trends. Customisable products are making headway in fashion. Nike offers Nike By You, which allows consumers to customise shoe colours and materials using 3D modelling, while rival Adidas offers several customisation options during checkout. These tools could also offer retailers valuable data to anticipate consumer trends. On the designer side, tools like CALA make it possible to quickly generate, prototype, and produce new designs, leveraging OpenAI’s DALL-E API. Vizcom has raised $5 million in funding for its tool that turns 2D sketches into 3D photorealistic renders.⁹

Large volumes of optimised product copy and imagery can be easily generated by AI, which can also translate, optimise for search engine optimisation (SEO), or A/B test content. AI can also quickly generate product images and 3D models. Google data reveals product offers with more than one image see 76% more impressions and 32% more clicks.¹⁰  Bloom, the product photo performance engine with $1.1 million in funding, tracks consumers’ on-site shopping behaviours to generate and deliver imagery with personalised models and environments.¹¹  NexTech, a 3D model supplier to Amazon, helps brands create Metaverse experiences and announced a breakthrough in fast, AI-powered 3D model texture creation in early 2023.¹²  Meanwhile, studies have shown that 60% of brands and retailers plan to invest in AI to automate product visual content online.¹³  Well-known players in e-commerce help smaller businesses customise their sites, including Google’s Product Studio, Shopify, which offers Shopify Magic, and eBay, with its Smart Store platform. Amazon now offers an integrated LLM tool for sellers to write product listings. Dedicated tools like Narrativa help automate the copywriting process for e-commerce, turning product specifications into unique product descriptions.

Personalised consumer journeys mean that now every shopper can have a unique site experience and custom-picked products, reducing shopper fatigue and boosting conversion. AI tools can analyse customer browsing, purchasing, and zero-party data to generate recommendations while honing predictive accuracy over time. Up-and-coming companies enabling e-commerce personalisation include Bloomreach, and Insider, with £274 million in funding.¹⁴  Algolia is a site search and discovery platform, and Adobe’s ‘Project Fast Filtered’ simplifies large-scale product navigation. On the retailer side, EDreams ODIGEO, one of Europe’s largest e-commerce firms, is working with Google Cloud to build tailored travel planning experiences. Stitch fix, an online personal styling service, openly uses generative AI for outfit selection, text composition, and feedback analysis at scale. Forbes data suggests that half of consumers agree they are comfortable using AI to improve purchase experiences; 61% trust AI to deliver them tailored promotions and deals.¹⁵

Generative AI could improve profit margins by optimising logistics

To cope with supply chain disruptions and shrinking profit margins, e-commerce leverages AI to streamline logistics across production, shipping, warehousing, inventory, and pricing. AI helps retailers implement dynamic, optimised pricing and procurement based on real-time markets or consumer behaviour, whether discounts or upsell opportunities. It can also help retailers manage supplier relationships and negotiate procurement contracts. A range of companies, like DLabs.AI, Retalon, and Alibaba Cloud, offer dynamic pricing and AI-powered promotions for consumers, often as part of broader personalisation strategies. Pactum AI enables major retailers, like Walmart, to automate supplier procurement negotiations for pricing optimisation across thousands of suppliers. 58% of top retailers already planned to implement AI pricing technology by the end of 2021.¹⁶

AI demand forecasting uses historical data, market trends, user sentiment data, and input from robots to help brands make informed manufacturing and inventory optimisation. Mass customisation through AI design tools could help retailers avoid costly over-stocking. Fashion group URBN uses AI-driven o9 Solutions tools for demand planning across its global brands, while H&M Group partnered with Google Cloud in 2020 to improve demand forecasting. AI-powered demand forecasting could cut supply chain network errors by 50% while speeding service.¹⁷  Walmart uses AI to streamline e-commerce fulfillment, while Amazon uses AI algorithms to predict stock demand. Dexory, a startup with $19 million in funding, offers robots and AI for inventory management.¹⁸  AI/ML-powered platforms for business operations include AWS Supply Chain, Google Cloud’s Vertex AI, and Microsoft’s Supply Chain Platform, plus services from Sentient Technologies, IBM, Oracle, and Dataiku. Alibaba Cloud offers AI-powered analytics and robotics for production, while Alibaba Group (though currently facing regulatory challenges) led the shift to mass e-commerce customisation as early as 2017.

AI can help optimise every part of supply chain logistics, from suppliers and warehousing to last-mile routing and delivery, helping to reduce costs, improve efficiency, and avoid delays. Customer-facing bots could provide real-time shipping updates. Some companies focus on AI tools specifically for logistics, training dedicated LLMs for industry use. Rippey.ai offers proprietary bots for forwarders that provide shipping quotes and track and book shipments. Microsoft offers an AI-driven assistant for integration in customer relationship management (CRM) and enterprise resource planning (ERP) systems. Industry data shows that 80% of logistics professionals either plan to use or already use AI for pricing, shipping, or customer support.¹⁹

Within product data and category management, AI can automate the organisation, tagging, and curation of extensive catalogues. It efficiently labels, categorises, and removes duplicates, accelerating processes and creating an intelligent taxonomy for improved search and recommendation algorithms. Investors may want to note players who offer solutions in this e-commerce niche, including Lily AI,²⁰  with over $39.5 million in funding, Crossing Minds²¹ , with $13.5 million in funding (including strategic investment from Shopify), Bloomreach²²  and Algolia²³ . Other companies of interest include Pixyle, which offers image recognition software for fashion retailers, and DotActiv, which works on automated product classification.

Companies shown are for illustrative purposes only as of 22/04/2024. It does not constitute investment research or financial analysis relating to transactions in financial instruments, nor does it constitute an offer to buy or sell any investments, products or services, and should not be considered as solicitation or investment, legal or tax advice, a recommendation for an investment strategy or a personalised recommendation to buy or sell securities.

Risk warning
No assurance can be given that our investment strategies will be successful. Investors can lose some or all of their capital invested. Our strategies are subject to specific risks including, but not limited to: equity; emerging markets; global investments; investments in small and micro capitalisation universe; investments in specific sectors or asset classes, volatility risk, liquidity risk, credit risk, counterparty risk, derivatives risk, legal risk, valuation risk, operational risk and risks related to the underlying assets. Some strategies may also involve leverage, which may increase the effect of market movements on the portfolio and may result in significant risk of losses.