CAPTCHAs Are Costing Retailers Customers — Heavy AI Users Are the First to Walk Away
Content Marketing Manager
Retailers heading into the final stretch of the end-of-year shopping season are facing a growing liability in their security stack: CAPTCHAs — those quick puzzles website visitors must solve to prove they’re human — are driving away the customers most sensitive to friction.
Our survey of 881 U.S. online shoppers reveals that when a CAPTCHA interrupts checkout, 18.7% give up after one failed attempt, and 3.9% abandon their cart immediately to shop elsewhere. Among heavy AI users, these figures jumped slightly up: 20% give up after one attempt, and 4.4% leave immediately.
Although well-intentioned, CAPTCHAs are costing sales and introducing friction at the exact moment shoppers expect seamless flow. At the same time, CAPTCHAs offer diminishing security value: modern automation can bypass or solve many common CAPTCHA formats, leaving retailers with the downsides of the feature but fewer of the protections they expect, and newer alternatives, like Dynamic Challenges, provide security without disrupting legitimate users. Retailers need adaptive, layered defenses against bots that don’t sacrifice their bottom line.
Key takeaways:
CAPTCHAs block revenue: When faced with a CAPTCHA during checkout, nearly 1 in 4 shoppers (23%) either abandon their cart after one failed attempt or leave to buy elsewhere immediately.
CAPTCHAs are weak against modern automation: Many AI-powered bots can bypass static challenges faster than humans.
Modern shoppers, especially heavy AI users, judge retailers on speed: 77% of heavy AI users say site speed directly increases their trust in a brand. They are 7 percentage points less patient than non-AI users at checkout.
Adaptive, invisible security is becoming a competitive requirement: The industry no longer needs to choose between security and user friction. Retailers can deploy layered, intelligent defenses that operate in real time without interrupting shoppers.
Note: In this study, we define heavy AI users as respondents who use AI tools daily or several times per week. Rare or non-AI users are respondents who have used AI once or twice in the past three months, or not at all. When we compared heavy AI users with rare or non-AI users, age and device preferences didn’t explain the difference. The demographic split was nearly identical across groups, and while heavy AI users shop slightly more often through mobile browsers instead of retailers’ apps, the gap wasn’t large enough to drive the behavioral divide.
Heavy AI users have different attitudes towards CAPTCHAs
CAPTCHA is intended to signal security, but our survey data shows it generates more negative than positive feelings. When faced with one on a retail site:
27% of all shoppers feel frustrated or distrustful. This increases to 29% among heavy AI users.
47% of all shoppers are neutral toward CAPTCHA, compared to 41% of heavy AI users.
Among some users, CAPTCHAs can generate reassurance: 22% of all shoppers say a CAPTCHA makes them feel reassured. Heavy AI users are the most likely to feel reassured, but also the most likely to feel frustrated or distrustful by CAPTCHAs.
Heavy AI Users Are Less Patient Shoppers
Beyond their reactions to CAPTCHAs, heavy AI users behave differently across the shopping journey. AI tools can deliver instant answers and adapt without delay. When people become accustomed to instant responses, our survey data shows that they carry that expectation into other digital experiences.
As a result, speed has become a baseline for whether a site feels trustworthy:
Of heavy AI users, 77% say a fast, responsive site increases their trust, compared with 56% of non-AI users.
On top of this, heavy AI users tend to be less patient with slow sites:
78% of heavy AI users will wait more than four seconds for a product page, compared with 87% of non-AI users.
At checkout, 43% of heavy AI users will wait more than 45 seconds for payment to process, versus 50% of non-AI users.
Why This Matters for Retailers
As AI adoption becomes the norm, the expectations of these users will increasingly define mainstream standards. For retailers, this means the friction created by CAPTCHAs is expected to grow, not shrink.
On top of this, retailers face an increasing level of spam from AI bots. According to Fastly’s Q2 2025 Threat Insights Report, AI fetcher bots generated traffic spikes that, in some cases, reached up to 39,000 requests per minute to a single site. These surges can overwhelm unprepared infrastructure even without malicious intent.
Under stress, many teams fall back on CAPTCHAs as a safety measure. However, CAPTCHAs interrupt real users while often doing little to address the scale or sophistication of modern bot traffic.
A Better Path: Adaptive, Invisible Validation
Retailers already know the stakes. Nearly one in four shoppers abandon checkout when a CAPTCHA appears, and heavy AI users walk away even faster. Instead of interrupting every shopper with a frustrating interactive challenge, Fastly’s adaptive security feature, Dynamic Challenges, verifies legitimate users invisibly. Because Dynamic Challenges are integrated with Private Access Tokens (PATs), validated users skip challenges entirely. This dramatically reduces the number of CAPTCHA puzzles real customers ever see.
If a CAPTCHA is costing you conversions, Fastly can help. Connect with our team to see how Dynamic Challenges keep real users shopping without the friction.
Methodology
This survey was conducted online between October 22 and November 2, 2025, and included 881 U.S.-based online shoppers. Respondents were grouped by self-reported AI usage frequency into two main categories:
Heavy AI users (use AI tools daily or several times per week)
Rare or non-AI users (used AI once or twice in the past three months, or never)
Generational breakdown showed consistent representation across user groups: Gen Z comprised 26.05% of heavy AI users and 26.56% of rare/non-users. Millennials were the largest group in both categories (45.41% and 45.81%). Gen X represented about 24% of each group. Boomers made up around 4%, with minimal variation.
All respondents reported recent online shopping activity. No participants selected “don’t shop online.” Survey results may reflect limitations common to self-reported data.