Search, Recommend, Reward: How AI personalization converts visitors into high-LTV shoppers

At Balkan eCommerce Summit 2026, Tudor Goicea presented a highly practical session focused on one of the biggest challenges facing modern eCommerce brands:

how to increase customer lifetime value in an environment where acquiring new customers is becoming increasingly expensive and difficult.

The presentation explored how AI-powered personalization can help online stores maximize the value of existing website traffic through three critical components:

• Search
• Recommendations
• Rewards

Together, these elements create what Tudor described as:
“the flywheel of lifetime value.”

The session also explored:

• changing consumer behavior,
• marketplace competition,
• AI-driven shopping experiences,
• and the future impact of agentic AI on eCommerce.

The Old eCommerce Growth Playbook No Longer Works

Tudor opened the session by explaining that the traditional eCommerce growth model that worked reliably in previous years is becoming increasingly ineffective.

For many brands, the old formula was relatively simple:

• buy traffic,
• drive users to the website,
• convert visitors into customers,
• scale acquisition.

Today, however, eCommerce businesses are being squeezed by multiple market forces simultaneously.

According to the presentation, three major pressures are reshaping online commerce:

• Rising customer acquisition costs
• Declining consumer confidence
• Increasing marketplace dominance

This creates a difficult environment where brands are:

• paying more to acquire customers,
• while customers themselves are becoming more cautious about spending.

Rising Acquisition Costs & Declining Consumer Confidence

One of the most important statistics shared during the session was:

• customer acquisition costs have increased by approximately 60% over the past five years.

At the same time:

• consumer confidence has dropped significantly.

This means eCommerce brands are now investing more money simply to achieve the same level of customer acquisition and conversion performance.

The presentation emphasized that this fundamentally changes how online stores should think about growth.

Instead of focusing only on acquiring more traffic, brands must become significantly better at:

• converting existing visitors,
• increasing order value,
• improving retention,
• and maximizing lifetime customer value.

Marketplace Dominance Is Increasing

Another major challenge discussed during the session was the growing dominance of large marketplaces across regional eCommerce markets.

Tudor referenced several major platforms including:

• eMAG
• Trendyol
• Temu
• Fashion Days
• GLAMI

These platforms increasingly dominate through:

• massive product assortments,
• strong logistics,
• aggressive pricing,
• large marketing budgets,
• and highly optimized user experiences.

Tudor explained that independent online stores are unlikely to outperform marketplaces in areas such as:

• delivery speed,
• assortment scale,
• or pricing power.

Instead, independent brands must compete through:

• customer experience,
• personalization,
• relevance,
• and retention.

The key is making every visitor count.

The Three Pillars of Lifetime Value

The core of the presentation focused on three major components of AI-driven personalization:

Search

Recommend

Reward

Tudor argued that many online stores manage to execute one or two of these effectively, but very few successfully integrate all three into a unified customer experience system.

When connected properly, these elements create a self-reinforcing personalization loop that increases:

• conversion rates,
• average order value,
• retention,
• and lifetime value.

Search: The Highest-Intent Users on Your Website

The presentation identified on-site search as one of the most valuable and underoptimized components of the eCommerce experience.

According to the data shared:

• approximately 69% of shoppers go directly to the search bar when visiting an online store.

These users typically demonstrate:

• high purchase intent,
• clear buying goals,
• and significantly stronger conversion potential.

Tudor explained that users interacting with search often convert:

• two to three times more effectively than average visitors.

This makes search one of the most important moments in the customer journey.

Why Search Experience Matters

The presentation compared eCommerce search functionality to the first interaction with a salesperson in a physical store.

If search fails:

• customers become frustrated,
• cannot find products quickly,
• and often leave the website entirely.

Modern AI-powered search systems should include:

• typo tolerance,
• semantic understanding,
• natural language interpretation,
• personalized ranking,
• and extremely fast response times.

One of the major behavioral shifts discussed was the evolution from:

• keyword-based search
  toward
• conversational search behavior.

Instead of searching for:
“black dress”

users increasingly search naturally using prompts such as:
“I’m looking for an evening dress for a wedding in July.”

This shift is being accelerated by widespread adoption of large language models and conversational AI systems.

Personalized Search & Real-Time Relevance

Tudor also emphasized the growing importance of personalized ranking.

Modern search systems can increasingly predict user intent based on:

• browsing behavior,
• purchase history,
• session activity,
• and customer preferences.

This allows brands to:

• prioritize relevant products,
• surface likely purchases faster,
• and reduce friction during discovery.

Speed is also critical.

According to the presentation:

• search results should ideally appear within 50 milliseconds,
• and definitely under 100 milliseconds.

Anything slower risks reducing engagement and increasing abandonment.

Recommendations: Growing Basket Size & Revenue

The second pillar of the personalization flywheel focused on product recommendations.

These include modules such as:

• Frequently Bought Together
• Recommended for You
• You May Also Like
• Similar Products

Recommendation systems primarily influence:

• average order value,
• product discovery,
• and basket expansion.

According to Tudor, most customers purchase:

• fewer than two items per order on average.

This creates a significant revenue gap between:

• what users buy
  and
• what they could buy.

AI-driven recommendation systems help close this gap.

Recommendation Engines Drive Measurable Growth

Several case studies were presented showing the measurable impact of recommendation systems.

Camalino

Users interacting with recommendations achieved:

• 5x higher conversion rates
• 3x higher average order value

Nikki Vibes

Recommendation-driven purchases resulted in:

• 25% of all orders containing recommended products
• 3x higher conversion rates
• 55% more items per basket

According to the presentation:

• recommendation-influenced revenue reached approximately 14%.

Combined together, search and recommendation systems can influence:

• 20–25% of total revenue.

Reward: Beyond Traditional Discounts

The third pillar of the framework focused on rewards.

Tudor explained that rewards extend far beyond traditional coupon systems.

Rewards may include:

• Free shipping
• Early access
• Gifts
• Personalized offers
• Dynamic discounts
• Retention incentives

The purpose of rewards is not simply to increase short-term conversions, but to:

• reduce hesitation,
• strengthen retention,
• and build long-term customer habits.

Retention Is Far More Valuable Than Acquisition

One of the most important insights from the session was the cost difference between:

• acquiring new customers
  and
• retaining existing ones.

According to the presentation:

• acquiring a new customer can cost between 5 and 25 times more than retaining an existing customer.

However, Tudor also warned against overusing generic discounts.

Permanent discounting:

• damages margins,
• weakens brand perception,
• and trains customers to wait for promotions.

Instead, reward systems should become:

• dynamic,
• behavior-driven,
• and context-sensitive.

Examples included:

• in-session incentives,
• reactivation campaigns,
• second-purchase encouragement,
• and loyalty-building rewards.

The Personalization Flywheel

The central concept of the presentation was the interconnected nature of:

• Search
• Recommend
• Reward

Each component continuously improves the others.

Search feeds Recommendations

Better understanding of user intent improves recommendation quality.

Recommendations feed Rewards

Behavioral insights help personalize incentives.

Rewards feed Search

Returning customers generate more behavioral data, improving future personalization.

This creates a self-reinforcing flywheel that continuously increases:

• relevance,
• customer engagement,
• and lifetime value.

Merchandising as the Control Layer

Tudor also introduced merchandising as the operational layer controlling the entire personalization system.

AI merchandising systems can:

• prioritize high-margin products,
• suppress out-of-stock items,
• reduce warehouse fragmentation,
• optimize campaign visibility,
• and protect profitability.

This ensures personalization does not simply maximize conversions, but also aligns with:

• operational efficiency,
• profitability,
• and inventory strategy.

Agentic AI & the Future of eCommerce

Toward the end of the session, Tudor briefly explored the growing impact of:
agentic AI.

He described two possible future scenarios.

Scenario 1: Buyer Agents Dominate

AI systems such as conversational shopping assistants increasingly control product discovery and purchasing decisions.

In this world, brands risk losing direct relationships with customers unless their infrastructure becomes machine-readable and discoverable.

This may require:

• Universal Commerce Protocols
• AI-readable product data
• Agent-compatible commerce infrastructure

Scenario 2: Brands Maintain Direct Experiences

Customers still visit brand websites, but interact through AI-powered on-site assistants and personalization systems.

In this case:

• internal personalization flywheels,
• AI search,
• recommendations,
• and contextual rewards

become critically important competitive advantages.

Aqurate’s AI Personalization Ecosystem

The presentation concluded with an overview of Aqurate and its AI personalization solutions.

The company focuses on:

• AI search
• Product recommendations
• Merchandising optimization
• Personalization systems

for eCommerce businesses across the Balkan region.

According to Tudor, stores using these systems have already achieved:

• measurable transaction growth,
  including:
• approximately 17% increase in transactions during early 2026.

Key Takeaways from the Session

The presentation delivered several important lessons for modern eCommerce brands:

• Rising acquisition costs make retention and personalization more important than ever
• Independent online stores cannot compete with marketplaces on scale alone
• On-site search is one of the highest-converting areas of eCommerce
• AI-powered recommendations significantly increase basket size and revenue
• Dynamic rewards improve both conversion and retention
• Personalization should function as a connected ecosystem, not isolated tools
• Lifetime value is becoming more important than one-time conversions
• AI-driven commerce will increasingly depend on machine-readable infrastructure and personalized customer experiences

Perhaps the strongest message from the session was that brands have already paid for the traffic reaching their websites – the real opportunity now lies in maximizing the value of every visit through smarter, AI-driven customer experiences.