Retailers Adjust to AI: What Matters in the New Normal of E-commerce?

The ongoing development of AI tools is having a profound impact on e-commerce. Consumers are increasingly using generative AI tools like ChatGPT to search for, select, and even purchase products, a development that affects every level of the e-commerce marketing funnel. Meanwhile, retailers are using AI tools to gather real-time public web data for purposes like dynamic pricing, demand forecasting, and inventory management.

Most importantly, these changes are happening quickly, and e-commerce retailers need to keep up. According to research, 67% of customers don’t think companies are reacting to their changing needs quickly enough. The peak e-commerce season of 2025, spanning from the lead-up to Black Friday to year-end holidays, is the first major test of how customers are using AI and how well retailers have adjusted and can leverage it.

From SEO to GEO

According to Adobe Analytics, traffic to retailers from generative AI tools like ChatGPT, Perplexity, and Claude jumped 1,200% from July 2024 to February 2025. Meanwhile, a survey found that 23% of shoppers plan to use chatbots and AI tools this holiday season, with that figure rising to over 42% among Gen Z and millennials. And AI’s impact now extends beyond product discovery, with OpenAI recently rolling out Instant Checkout for ChatGPT, enabling shoppers to make purchases without leaving the tool. It is currently available on Etsy and some Shopify stores.

These developments mean e-commerce retailers must rethink content, marketing, and sales. The consultancy Bain estimates that a significant majority of consumers now rely on zero-click results (where answers are provided by an AI overview instead of visiting a website) in 40% of their searches. This means less traffic to retailers’ websites, although conversion rates of those who do arrive on sites from AI sources are higher.

In this context, while traditional search engine optimization (SEO) tactics remain pertinent, the trajectory is clearly towards the use of generative AI for shopping, hence the emergence of Generative Engine Optimization (GEO). GEO presents e-commerce merchants with a range of novel challenges. Large Language Models (LLMs) that power generative AI tools are trained to evaluate reputation, credibility, and trustworthiness when analyzing brands. Therefore, retailers need to work hard to build their credibility, especially by gaining reviews or recommendations from well-respected external sources.

Descriptive product queries

Another GEO factor to grapple with is the different way customers form queries when using generative AI tools. According to OpenAI, almost half of all queries use “asking” patterns. The Chief Information and Product Officer at US retail giant Target recently claimed 25% of search requests made on their platform are now considered ‘descriptive queries’ that are complex and sophisticated.

Whereas on search engines a customer might search for a “slim-fit pink shirt,” the same query on an AI tool might be “Slim-fit pink shirts for business casual events.” For descriptive queries like this, product descriptions need to be adapted. For example, on product pages, the GEO best practice suggests including much more product description in the form of precisely written, detailed FAQs. This enables AI crawlers to easily identify which enquiries your product would be a good fit for.

Digital sandboxes for GEO

In an ironic twist, AI is being used to help with content analysis and GEO. Researchers at Columbia Business School are using large language models (LLMs) to create “digital twins” that mirror human behavior. When a specific product is input, the LLM generates a digital twin with a shopper persona, including name, age, occupation, and preferences. This twin then conducts relevant searches on ChatGPT to see how prominently the company’s product is listed. Companies can then leverage generative AI to adjust how their products are described and presented, based on the findings of these digital twins.

A ‘digital sandbox’ approach like this can be a productive way for e-commerce companies to conduct GEO, but it is not without risks. AI agents have their own biases, which may affect how they perform and behave. Nevertheless, these approaches provide a potential way forward in e-commerce intelligence.

AI-powered data collection

The marketing funnel is just one element of e-commerce being disrupted by AI. A potentially more important area is business intelligence (BI), a broad term describing the collection and use of data to generate insights that improve strategy and operations. For effective BI, e-commerce companies need reliable, up-to-date datasets, including external data. AI is now playing an important role in collecting competitive data.

The practice of extracting public web data, such as prices and product descriptions, has been a mainstay of e-commerce competition for years. Now AI is streamlining it. AI-enabled tools can be prompted using natural language, meaning no coding is required, and engineers do not need to spend hours building a full data collection pipeline. AI can also gather and filter suitable URLs for scraping, for example, by finding all product pages for a particular category on a competitor’s website.

With the rise of AI-powered shopping assistants, e-commerce companies will also be more inclined to collect data points from one another that appear only after specific actions are completed, for example, the final checkout price.

Demand forecasting and reacting in real time

With an array of real-time data collected, from competitor pricing to inventory, retailers can adjust their pricing or marketing immediately and provide the best offers to the customers.

Dynamic pricing is one of the most important and popular BI functions retailers can use, and according to a recent survey, 61% of retailers in Europe make use of it. However, the same survey found that less than 15% use algorithms or AI for this purpose, revealing an opportunity. Leveraging the latest data on competitive pricing, LLMs can be trained to automatically adjust pricing, which is especially useful during peak periods such as the holiday season.

AI can use data on customer demand and stock levels to forecast future demand. This can bring multiple benefits. Deloitte Digital has highlighted how retailers can use AI to monitor their own stock, manage inventory, and place orders dynamically. Additionally, AI can help analyze data gathered across the web to understand how a brand is viewed, providing strategic-level insights.

Open to opportunity

While AI is disrupting the e-commerce marketing funnel, it is also creating new opportunities. It can be leveraged to analyze and create geo-optimized content. It is powering efforts to gather valuable real-time public web data. AI is also adding value in analyzing data to make decisions on pricing, inventory, and strategy. And that is not to mention other potential uses, such as enhanced customer support.

Something as disruptive as AI is always scary, especially for retailers approaching their main sales period. However, those that remain open to the opportunities AI is creating can do more than just survive. They can thrive.