# AI Agents Seek To Automate and Personalize Purchase Journeys | Capterra > Emerging AI agents can automate the often arduous task of searching for products online. Data security, hallucinations, and AI-washing pose adoption barriers. Source: https://www.capterra.com/resources/retail-ai-agents --- Small Business Customer ServiceIT & Software Development # Online Shopping is Due for an Overhaul—Can AI Agents Help? Written by: Molly Burke Molly BurkeAuthor Senior Specialist Analyst Experience I have been writing content for Capterra since April 2022. I cover technology trends in retail, hospitality, and custome... [See bio & all articles](https://www.capterra.com/resources/author/mburke/) Published June 12, 2024 8 min read Table of Contents - [Shopping online has become arduous](#shopping-online-has-become-arduous) - [AI agents aim to take the legwork out of online shopping](#emerging-ai-agents-aim-to-take-the-legwork-out-of-online-shopping) - [Retailers may not get ROI from AI agents just yet](#much-hyped-ai-agents-may-not-provide-roi-to-retailers-just-yet) - [Retailers can improve customers’ experience now](#retailers-can-improve-customers-experience-now-while-waiting-for-ai-agents-to-improve) ## AI agents aim to provide a personalized and automated search experience. According to many consumers, the abundance of choice has actually made shopping online more inconvenient. Capterra’s 2024 Elusive Online Consumer Survey of over 5,500 consumers across 12 countries reveals that the large amount of inaccurate, irrelevant, and sponsored search results is a major challenge consumers face when shopping online.[\*](#methodology) Emerging conversational AI agents are poised to solve issues related to online product search by acting as virtual personal shoppers capable of delivering powerful, personalized automation. However, retail business leaders should take advantage of existing technology to solve search-related issues affecting their audiences, while _also_ preparing for the growing influence of AI agents over online purchase decisions. Key insights - Over half of global consumers say dealing with too many sponsored results and inaccurate or irrelevant results when searching for products online is a challenge.  - 23% of global consumers find it challenging to know which search terms or keywords to use when shopping online. - 84% of consumers encounter issues when using online search filters, including incorrectly applied filters, a lack of specificity, or having too few filters to meaningfully narrow results. ## Shopping online has become arduous  Today’s online shopper has to do a fair bit of investigative work to find the right product. First, they have to know which search terms to use, which can differ across channels as retailers scramble to keep up with social media’s growing influence on purchase journeys.  Nearly a quarter (23%) of consumers struggle with knowing which keywords to use on search engines, retailer websites, and eCommerce marketplaces. It’s even more challenging for consumers who get inspiration from social media—they often have to translate viral trends into retailers’ product naming conventions.   Over half of global consumers then have issues making sense of search results, which they say are cluttered with sponsored products, inaccuracies, and irrelevant suggestions. Nearly a third (31%) say they receive too many search results altogether. They then deal with the following challenges when attempting to narrow down their options using search filters: - 43% say search filters are incorrectly applied to products - 42% say search filters are not specific enough - 31% say there are too few search filters to meaningfully whittle down their options After they’ve managed to create a shortlist of products, consumers then have to do their due diligence in the review section to verify the quality of each product. They check a variety of critical review attributes, including the number of reviews both positive and negative, user-generated images of the products in action, and the star rating.  The sheer amount of time and attention needed to sift through the vast sea of options impacts the customer experience, creating frustration and exhaustion. It also privileges larger retailers, who tend to show up near the top of online search results, making it difficult for smaller retailers to be discovered by online customers. ## Emerging AI agents aim to take the legwork out of online shopping AI agents seek to lighten the load by automating much of the manual work today’s online shoppers do when browsing products. **What is an AI agent?** An AI agent is software that can perceive information about its digital or physical environment, make decisions autonomously, and take action. They are designed to interact with the internet on behalf of human users, like robot virtual assistants. In a retail context, they resemble personal shoppers: They collect information about what a customer wants and returns a list of products or services that would make a good fit, with helpful information and analytics to guide the customer’s purchase decision. The agent handles all the work so customers don’t have to dream up keywords, fiddle with filters, or process troves of reviews for the high notes.  Some AI agents work entirely behind-the-scenes, generating code and accessing APIs, while others are being trained to navigate the internet like people do (but faster), using clicks and keystrokes.[\[1\]](#sources) On the cutting edge of development are “multimodal” AI agents—chatbots that can accept and generate text, sounds, and images at humanlike speed. They lack the latency that creates that uncanny-valley feeling you often get when talking to a machine.[\[2\]](#sources)  In essence, retail AI agents would replicate the experience of talking to a store associate with deep knowledge of products and a knack for personalized assistance. ## Much-hyped AI agents may not provide ROI to retailers just yet Experts predict that by 2026, half of customer service and support organizations will implement customer-facing generative AI agents.[\[3\]](#sources) But AI agents’ success in retail largely depends on whether they actually work as promised, as well as whether shoppers are willing to share their personal information with AI providers, who collectively have a lousy track record for data security.[\[4\]](#sources) ### Retailers should beware costly hallucinations AIs are prone to creating and sharing misinformation, a phenomenon known as “AI hallucination.” The issue has already resulted in costly scandals for some large companies, such as when Air Canada was forced to refund a customer after its generative AI chatbot lied about the company’s bereavement policy.[\[5\]](#sources)  Frequent hallucinations could render AI agents effectively useless, causing low use among customers and costing retailers more money than they generate. An AI agent miscommunicating a retailer’s return policy, suggesting products that don’t exist, or failing to understand a user’s query could lead to a spectrum of headaches for businesses, ranging from an influx of customer service helpdesk requests to, in Air Canada’s case, a lawsuit. ### Data privacy and security are top AI-related issues for businesses and consumers  Then, there are the issues of data privacy and data security. Previous Capterra research on U.S. consumers shows that while many are intrigued about how AI could help them shop online, the majority doubt their [data will be handled responsibly by AI tools](https://www.capterra.com/resources/retail-generative-ai/) and their providers, and they don’t want AI providers to sell their data to advertisers. To that end, most are reluctant to share personal information that could lead to personalized results, such as details about their lifestyle or images of themselves.  Consumers aren’t alone in their concern over sensitive data—businesses regard data leaks, not hallucinations, as the biggest threat posed by AI. According to recent research, nearly half (45%) of companies surveyed about AI encountered issues with exposed data when they tried implementing AI software.[\[6\]](#sources) ### AI-washing harms businesses and consumers One proposed solution to issues around hallucination and data security is “human-in-the-loop” AI, wherein human supervisors are involved in various points of an AI’s functioning, influencing the AI’s behavior to mitigate risks.[\[7\]](#sources) In retail, a human-in-the-loop strategy could include assigning employees to validate an AI agent’s output before it has a chance to pass on misinformation to the customer. Similarly, businesses could assign humans to conduct regular security audits to catch vulnerabilities in the AI before they lead to full-blown data leaks.  That brings us to “AI washing,” an emerging buzzword describing the discrepancy between how autonomous today’s AI tools really are, versus how they are marketed to both business-to-business (B2B) and business-to-client (B2C) customers. AI washing takes the idea of a human in the loop to the extreme—the humans are actually performing the bulk of the work on behalf of the incompetent or error-prone AI.  Several such high-profile cases have recently come to light, wherein remote workers were performing most or all of the work behind the scenes of “AI-driven” drive-through machines, cashierless grocery stores, and content moderation algorithms.[\[8\]](#sources)  AI washing is a risk to both consumers and businesses: Its false promises erode consumer trust in AI, while leading retailers to make [unsatisfying software purchases](https://www.capterra.com/resources/retail-trends/). ## Retailers can improve customers’ experience now while waiting for AI agents to improve  Software experts already consider generative AI to be highly useful for content generation, conversational user interfaces, and knowledge discovery.[\[9\]](#sources) Businesses can certainly leverage those capabilities to enhance customer experience on their websites. That said, the bottom line is that all-in-one AI agents are currently in a rudimentary phase of development, and still come with significant risks that require human oversight.[\[10\]](#sources) At worst, an AI agent could actually cost retailers money in the form of legal fees and reputational damage.  Retailers should carefully evaluate potential investments into AI tools, prioritizing low-risk options such as customer-facing chatbots with human-in-the-loop oversight and predefined prompt input.[\[3\]](#sources) More sophisticated AI agents that make fewer mistakes and have better upselling efficacy are on the horizon. In the meantime, retailers don’t need to wait to enhance customers’ experience—they can improve their product discoverability and site search functionality using existing tools and resources. Much of the following software includes AI features and automation that lighten the user’s workload. - Use [social media monitoring software](https://www.capterra.com/social-media-monitoring-software/) to discover the trends and keywords associated with products similar to those you sell. Incorporate those keywords into product metadata, so customers can find your product on search engines and your own eCommerce website.  - Redesign your eCommerce website with [website builder software](https://www.capterra.com/website-builder-software/) to make the search bar more visible to site visitors. It’s a small change that can make a surprising difference in how frequently customers search your site.  - Emulate the level of granularity that leading retailers offer with faceted search—in other words, let them narrow their options by multiple, category-specific attributes, such as sleeve length, dimensions, or materials. The specificity of faceted search can help customers find products they’ll like on your site more quickly. [Enterprise search software](https://www.capterra.com/enterprise-search-software/) can help.  - Display data visualizations based on customer reviews data, such as comfort rating or recommendation rating, to give customers a snapshot of how products perform. [Review management software](https://www.capterra.com/review-management-software/) helps businesses request and analyze customer feedback.  If you’re not sure where to start with improving site search, hiring experts can help. You can browse Capterra’s list of top user experience design agencies to find one that’s right for your business. ## Capterra's 2026 Software Buying Trends Report ### Download our 2026 Software Buying Trends Report to see how successful software adopters avoid disappointment and how your business can, too. * * * **Survey methodology** \*Capterra's 2024 Elusive Online Consumer Survey was conducted online in April 2024 among 5,585 respondents in the U.S. (n: 500), Canada (n: 500), Brazil (n: 497), Mexico (n: 470), the U.K. (n: 499), France (n: 271), Italy (n: 496), Germany (n: 496), Spain (n: 359), Australia (n: 497), India (n: 500), and Japan (n: 500). The goal of the study was to learn about how today's online consumer shops. Respondents were screened to have shopped online several times a month or more often. Sources 1. [How ‘A.I. Agents’ That Roam the Internet Could One Day Replace Workers](https://www.nytimes.com/2023/10/16/technology/ai-agents-workers-replace.html), The New York Times 2. [OpenAI Unveils New ChatGPT That Listens, Looks and Talks](https://www.nytimes.com/2024/05/13/technology/openai-chatgpt-app.html), The New York Times 3. [Innovation Insight: Generative AI Chatbot to Improve CX and Agent Productivity](https://www.gartner.com/document/5372763?ref=solrAll&refval=411911344&), Gartner 4. [ChatGPT is leaking users' passwords, report finds](https://www.thestreet.com/technology/chatgpt-sam-altman-artificial-intelligence-privacy-ethics-passwords), The Street 5. [Air Canada Has to Honor a Refund Policy Its Chatbot Made Up](https://www.wired.com/story/air-canada-chatbot-refund-policy/), Wired 6. [The biggest risk corporations see in gen AI usage isn’t hallucinations](https://www.cnbc.com/2024/05/16/the-no-1-risk-companies-see-in-gen-ai-usage-isnt-hallucinations.html), CNBC 7. [Design and Implement Human-in-the-Loop Interfaces for Control, Performance and Transparency of AI](https://www.gartner.com/document/4018343), Gartner 8. [A Brief History of Automatons That Were Actually People](https://www.scientificamerican.com/article/is-there-a-human-hiding-behind-that-robot-or-ai/), Scientific American 9. [When Not to Use Generative AI](https://www.gartner.com/en/articles/when-not-to-use-generative-ai?), Gartner 10. [Innovation Insight: AI Agents](https://www.gartner.com/doc/5332663?ref=solrAll&refval=411911344&), Gartner * * * Looking for Retail Management Systems software?Check out Capterra's list of the [best Retail Management Systems software](https://www.capterra.com/retail-management-systems-software/) solutions. ### Was this article helpful? * * * ## About the Author [### Molly Burke](https://www.capterra.com/resources/author/mburke/) Molly Burke is a senior analyst and writer for Capterra. She covers customer experience and marketing in the retail and restaurant industries, with a focus on how emerging technology is transforming the way everyday people shop. Her insights on generative AI, social media, and other tech trends have been featured in The New York Times, Vogue, BBC, CNBC, Forbes, and the Financial Times, among other publications. ### RELATED READING - [Free CRM Software vs. Paid: Features, Costs, and Comparison](https://www.capterra.com/resources/free-vs-paid-software-for-crm-guide/) - [Accounting Trends in 2026: How AI Is Changing Work, Skills, and Strategy](https://www.capterra.com/resources/accounting-trends-ai-software-changing-work/) - [Bookkeeper vs. Accounts Payable Software: Which One Should You Choose?](https://www.capterra.com/resources/category-compare-bookkeeper-vs-accounts-payable-software/) - [How to Unlock Real AI Benefits in Sales and Marketing](https://www.capterra.com/resources/ai-benefits-sales-marketing-software/) - [How To Build the Right Project Management Tech Stack for Your Business](https://www.capterra.com/resources/project-management-tech-stack/) - [5 Key Help Desk Software Features With Top Products That Offer Them](https://www.capterra.com/resources/help-desk-software-key-features/) - [Why You Need Project Management Software With ADP Integration](https://www.capterra.com/resources/project-management-software-with-adp-integration/) - [Help Desk Software for IT Support: Features and Benefits 1,300+ Users Value](https://www.capterra.com/resources/it-help-desk-software-features-benefits/) - [AI in Talent Management Software: The Bridge Between Recruiting and Retention](https://www.capterra.com/resources/ai-talent-management-software/)