How to protect customer journeys as AI changes holiday shopping

Holiday peak preparation often focuses on one question:

How many users can the platform support?

That matters, but protecting critical journeys requires more than a headline traffic target. It means testing realistic customer behaviour before peak and maintaining visibility once live trading begins.

As AI-assisted discovery changes how people research products, enter retail websites and move towards purchase, retailers may also need to rethink the assumptions behind peak preparation.

In this article, we look at what those changes could mean for load testing before peak and journey monitoring during it.

How AI-assisted customer journeys are changing traffic profiles

As explored in Part 1, AI-assisted discovery means more shoppers may arrive on a retailer’s website after already comparing products, prices and reviews elsewhere.

They may enter directly through a product page and move more quickly towards stock checks, basket updates, promotions and payment.

This can change where pressure builds across the platform. The same number of visitors will create a very different load profile if more of them arrive ready to purchase rather than browse.

Peak preparation should therefore reflect not only how many customers are expected, but the journeys they are likely to take once they arrive.

Review the journeys in your load test

Start by identifying the journeys that matter most to customers and the business.

These may include:

• browsing and search

• direct product entry

• login and account access

• stock checks

• promotional codes

• basket updates

• guest and registered checkout

• different payment methods

• delivery and collection options

The traffic mix is important.

A test dominated by browsing may place too little pressure on stock, basket, promotion and payment services. A test made up entirely of checkout journeys would be equally unrealistic.

The aim should be to recreate a credible balance of activity.

Include realistic paths and drop-offs

Not every customer completes every journey.

Some browse several categories. Others arrive directly on a product page. Some leave after checking stock, while others abandon their basket or payment.

Including these different paths and drop-off points creates a more realistic load profile.

It can also reveal where pressure begins to appear as activity moves from browsing towards more transactional steps.

Test where pressure appears across the journey

Critical customer journeys often rely on several systems and third-party suppliers.

A checkout may depend on stock, promotions, payment, delivery and order-management services. Each of these can respond differently as demand increases.

The storefront may remain available while stock checks slow down, promotional codes begin to fail or payment provider responses become intermittent.

Load Testing should therefore show more than whether the final journey completed. It should help identify:

• which step begins to slow down first

• whether errors affect all customers or only some journeys or customers

• which service or integration becomes a bottleneck

• whether performance recovers as expected

• how degradation affects the overall customer experience

This gives teams clearer evidence of where action is needed before peak, rather than simply confirming that the site remained online.

Continue assurance once peak begins

Load testing provides evidence before peak, but live conditions will still introduce the unexpected.

Campaigns go live. Promotions change. Releases continue. Third-party services behave differently. Real customers do not always follow the predicted paths.

That is why User Journey Monitoring remains important throughout peak.

Customers arriving through AI-assisted research may already be ready to buy and may have alternative retailers in mind. A slow or broken journey could make it easier for them to purchase elsewhere.

Monitoring critical journeys from outside the organisation helps answer:

Are customers still able to search, log in, add to basket and complete checkout right now?

Load testing and monitoring answer different but connected questions:

Can critical journeys cope with expected demand?

Are those journeys continuing to work once peak trading begins?

Together, they provide a clearer view of where risk exists before peak and whether customers are still receiving the experience they expect during it.

Protect critical journeys before and after peak

Explore how our User Journey based Load Testing and User Journey Monitoring services can help protect critical customer journeys before and during holiday peak.