Site search vs. ChatGPT: The future of B2B ecommerce unveiled

Jul 7th 2023 ecommerce

On Wednesday June 28th, Algolia joined seasoned industry experts brought together by MasterB2B to debate the key issues in B2B ecommerce. Hosted by renowned industry experts, Brian Beck and Andy Hoar, this insightful webinar aimed to explore whether ChatGPT, the fastest-growing app ever, might potentially outperform traditional site search methods in the B2B eCommerce sector.

The webinar gathered experts from diverse fields, including representatives from Genuine Parts Company and Algolia to defend traditional site search. Meanwhile, advocates for ChatGPT were professionals from Leica Biosystems and HP, who brought innovative perspectives from their respective sectors.

Round 1: Why site search is winning the war for personalization

Both teams acknowledged the importance of personalization in eCommerce experiences. Team ChatGPT argued that the AI model is more adept at understanding user intent, responding to user prompts and even offering opinions when asked. Plus it can be trained to interact with users in highly personalized ways. 

But in the context of e-commerce, personalization means understanding the right content within the right context. It means finding the specific product an individual user is looking for, with all the B2B centered specifications, and that takes a contextual understanding that is missing from ChatGPT. It lacks this B2B context. 

Site search, on the other hand, is inherently designed to capture behavioral data and user preferences using AI, already providing effective personalization without requiring additional input from users. For example, one Algolia customer, MSA Safety, sells to 4 different industries, so if you’re looking for a safety helmet – you’re going to want very different results if you work in construction than if you work in fire safety. A great search experience needs to understand the differences in specs, industry standards and compliance requirements, and most importantly context.

Generative AI models like ChatGPT lack access to data sources like PIM, CDP, or even buyer history and behavioral data, and are therefore more focused on handling the singular query experience versus optimizing the experience and personalizing it across the customer journey as site search solutions can do. 

At the same time, site search solutions need to respond to the broader trends like conversational search for the ease of use and a human-like experience. Those who don’t incorporate a solution to address will be left behind. That’s why Algolia is proud to be the only search provider on the market to incorporate both keyword and semantic search for each query powered by AI that can return the most relevant results under 20ms even when working on catalogs with tens of millions of product SKUs and various combinations.

Round 2: AI’s limitations in handling complex queries

The debate progressed to discussing whether generative AI, like ChatGPT, could handle various search queries. Team ChatGPT claimed that ChatGPT could process natural language and keyword searches effectively. 

But understanding the nuances makes or breaks the customer experience. Can ChatGPT handle specialized search queries, like product codes or proprietary descriptors? And how it could deliver specific requirements for B2B eCommerce, such as dimensions, units, filters, and autocomplete suggestions? Understanding these nuances requires access to private data which ChatGPT does not have. 

Imagine a customer is looking for a specific type of steel beam. The user might use very specific and nuanced search terms, such as material, dimensions, variance, temperature tolerance, industry certifications, and product nomenclature. A robust site search solution can parse this complex query, understand the context, and accurately provide the desired results. It can sift through thousands of product listings, apply multiple filters all at once to present the most relevant items. It can do so because the original index was built with all of these nuances captured already.

Site search solutions also leverage machine learning and other advanced AI technologies to continually refine and improve their understanding of user’s behavior, search patterns, and preferences. This means that the more a user interacts with the site search, the better it becomes at delivering personalized and nuanced shopping experiences.

Round 3: Conversational Commerce: who will get there first?

The final round tackled which tool would pioneer “conversational commerce”. Team ChatGPT argued that ChatGPT already delivers on the conversational aspect and maintained that their model’s advanced conversational abilities could easily integrate with various eCommerce platforms. 

But how many B2B buyers are interested in having a lengthy prompt-generated conversation? Team site search asserted that the commerce aspect is more foundational than the conversational piece. In B2B, buyers are mission-driven and know what they want. They need efficiency in being able to find the right products fast. Furthermore, generative AI is intended to mimic creative thinking and produce original results whereas site search is focused on relevance, specificity and precision.

Ultimately we need to be on a new path that incorporates the best of both worlds. The most advanced site search platforms, like Algolia, are already adept at understanding user intent and could easily incorporate a conversational interface. That’s why Algolia invested in building out search experiences for the future of ecommerce with its end to end AI platform and its NeuralSearch product, combining keyword and semantic search in a single API using a single index, customizable for your business needs – a value offering that no other competitor can match today, and our customers are eager to take it to the next level. 

For example, one customer wants to use GPT 4 as a conversational agent to respond to client questions regarding equipment that they sell. This conversational agent  would retrieve products and associated specifications, facets, and various attributes from their indices that are running on Algolia NeuralSearch as a way to further simplify and enhance the buying journey. This is the vision of the next generational search experience with conversational commerce that Algolia is working on.

In the rapidly evolving digital landscape, businesses need to continuously evaluate and adapt their eCommerce strategies. For businesses evaluating a search solution in the market, here are a few questions to help you better understand what’s under the hood:

• How is AI applied?
• Does semantic search alongside keyword search capability for each single query?
• Does the algorithm learn with in-session user engagement data?
• Can the tool provide clear visibility to the business users with no-code controls for dynamic tuning?
• And if there’s a claim of using a ChatGPT integration, understand first how they are tackling the security and proprietary data concerns of B2B.

The insights shared in this webinar underlined this imperative and provided a nuanced understanding of two powerful tools, generative AI and traditional site search, for enhancing B2B eCommerce experiences. You can watch the full webinar recording here.

If you need even more convincing, check out this post on why ChatGPT won’t replace search anytime soon.