By Sanjay Mehta, Head of Industry, Commerce at Lucidworks
The emergence of large language models demonstrates one of the most important functionalities of AI – returning information, products, assets from a large dataset. Search is particularly important in ecommerce where companies have a ton of SKUs, different types of content like videos, reviews, and blogs, and customer support FAQs. IMRG data shows that the website users who use onsite search convert at 3x those who don’t.
Search matters.
Most ecommerce users enter queries of three or fewer words, perhaps guessing at how the search engine works, or giving up and moving over to the browse experience. By training users to learn to search using metadata terms (for example, “women’s trainers white”) rather than natural language (women’s white trainers suitable for the office), we encourage ambiguity in search queries. The user wastes time thinking about the “right” way to search and misses out on a more explicit expression of what they are trying to accomplish.
Another aspect of today’s search experience is that it depends on implicit signals to predict the user’s general goal. Is this browse related to the same goal as the previous search? Is the user in research mode or ready to purchase? It is difficult to predict the user’s transition from one goal (new clothes for the office) to another (a gift for my mum). Can it interpret slang, e.g., attire vs threads vs clothes? Also, retailers should try to avoid ranking search results the same regardless of the user. Context and unique user history matters.
Search is complicated.
With the arrival of generative AI technologies like Bard and ChatGPT, a lot of companies see a potential quick fix for search. But to create excellent customer journeys, retailers can use large language models alongside personalised experiences and domain-specific knowledge.
The Commerce Industry’s Plans to Invest in Generative AI
Our own research at Lucidworks has made something very clear – AI, and especially generative AI enabled by LLMs – is an important element of the future of ecommerce. 100% of respondents to our search relevance survey ranked search relevance as highly important, with 96% saying it is difficult to deliver. A whopping 88% believe AI will be important in delivering relevant search in the future.
Practical Applications for Artificial Intelligence and Generative AI
Let’s put this in practical terms. These are some of the applications of generative AI and large language models, and the key benefits they provide when deployed efficiently.
1. Expand AI Insights Beyond Customer Experience
Global business leaders are predominantly directing their investments in AI toward three key areas: improved customer experience, automation and efficiency, and overall business operations. AI in the retail sector often centres around enhancing customer experiences through smarter search and shopping features, driven by AI-powered visual and natural language searching and intelligent digital assistants. While this customer-centric focus is crucial, it shouldn’t overshadow the potential of AI to optimise other aspects of a business.
AI algorithms can analyse complex shopping and behavioural patterns, revealing how customers discover the business and begin their buying journeys. Insights into successful buying paths, underperforming products, search result discrepancies, and abandonment rates can be harnessed for corrective actions. For instance, providing incentives for low-performing products, refining SEO strategies and product recommendations, addressing design flaws, promoting new product combinations, and diversifying search results.
Taking a comprehensive approach, AI can boost efficiency and profitability across the entire business landscape, encompassing marketing, finance, product design, and manufacturing. By leveraging AI insights beyond customer-facing aspects, retailers can unlock significant opportunities for growth and optimisation.
2. Turn Limited Chatbots into Smart, Conversational Virtual Assistants
The recent rise of ChatGPT has significantly raised people’s expectations when engaging with search engines or chatbots. To meet these elevated standards, many businesses are actively using large language model (LLM) AI technology into their online help, search, and digital assistant functions. The future demands that customers can express their needs and problems in natural, conversational language and receive relevant, relatable answers and recommendations. Failure to provide this experience may result in dissatisfied customers seeking alternatives.
However, transforming rigid chatbots into smart, conversational digital assistants requires more than simply integrating a large language model like ChatGPT. It involves refining the LLM-driven experience to align with the specific business and customer intents. This means tailoring virtual assistants to suit particular domains, such as a virtual stylist in luxury fashion, a virtual chef in the grocery sector, or a virtual tradesman in home improvement. By doing so, the dialogue with these assistants becomes contextually relevant and personalised to the individual shopper.
3. Introduce New Forms of Search
Similar to seeking help in a hardware store, where visual cues and direct communication aid in finding the right part, online shopping experiences also benefit from multi-modal searches. While traditional text-based search has its limitations, combining images, voice inputs, text descriptions, and other elements enhances the search process, making it more effective and user-friendly. AI has advanced to the point where it can accurately simulate visual representations, allowing customers to see how products like paint colours, furniture, or clothing items will look in specific settings or on particular individuals.
As AI continues to evolve, we can expect multimodal elements to become a standard feature in modern, competitive search experiences. This expansion will extend to augmented reality (AR) and virtual reality (VR) environments, integrating chat rooms, shopping experiences, and more. The future of search lies in catering to customers’ diverse communication preferences, ultimately providing a seamless and immersive experience that empowers them to make informed decisions.
4. Endless Other Applications
Here’s what gets us most excited about the potential of large language models:
- Hybrid search and LLMs form a “virtuous cycle,” grounding responses in fact
- LLMs improve semantic models
- Retrieval augmented generation grounds LLM responses in fact
- Conversational discovery
- LLMs enable personalised, information dense responses
- Document summaries
- Explanations
- Suggested filters
- Revenue-driving recommendations and personalisation
The power of Generative AI
The emergence of Large Language Models (LLMs) has unveiled the powerful functionality of AI in returning information, products, and assets from vast datasets. In ecommerce, efficient search capabilities can significantly impact conversion rates and customer satisfaction. By leveraging AI-driven insights into customer behaviour and preferences, businesses can optimise search algorithms, refine product recommendations, and address underperforming areas, leading to increased efficiency and profitability across the entire enterprise.
The integration of generative AI technologies, such as Bard and ChatGPT, presents a promising avenue for creating smart, conversational virtual assistants that cater to customers’ diverse communication preferences. By tailoring these virtual assistants to specific domains and incorporating multi-modal search capabilities, businesses can provide personalised experiences that resonate with individual shoppers, ultimately fostering loyalty and driving revenue growth.
As the industry looks forward, the potential of large language models in search is vast and exciting, offering possibilities for hybrid search, improved semantic models, conversational discovery, and more. By embracing AI and LLMs as integral components of their future strategies, commerce companies can stay ahead of the curve, delivering cutting-edge search experiences that redefine ecommerce in the digital era.
Explore more: Download the world’s largest study on Generative AI Usage
Published 05/09/2023
