The First Job of a Shopping Agent Isn’t Just Discovery. It’s Preventing Regret

Why It Matters

Reducing post‑purchase regret directly cuts the $890 billion loss, improving margins and customer loyalty. Retailers that shift focus to keep rate and confidence moments will gain a competitive edge in apparel and footwear.

Key Takeaways

• Retail returns hit $890 billion in 2024, 16.9% of sales
• 51% of Gen Z shoppers “bracket” items, buying multiples to return
• “Keep rate” metric gauges whether shoppers retain purchased items
• Mapping confidence moments—fit, fabric, comfort—reduces post‑purchase regret
• True Fit advocates a crawl‑walk‑run plan to improve agent trust

Pulse Analysis

The surge of AI‑driven shopping agents has reshaped how retailers surface products, yet the real battle lies beyond discovery. With the National Retail Federation projecting $890 billion in returns for 2024—roughly one‑sixth of total sales—retailers are confronting a cost structure that erodes profit margins. Gen Z shoppers amplify the problem, with more than half engaging in “bracketing,” purchasing multiple sizes or styles only to return the excess. This behavior underscores a confidence gap: shoppers are willing to buy, but they lack assurance that the item will meet their expectations.

Enter the "keep rate" metric, a shift from measuring intent to measuring validated outcomes. By tracking the proportion of items that remain in the consumer’s wardrobe, retailers gain a concrete signal of confidence at the moment of purchase. Mapping "confidence moments"—fit, fabric feel, comfort, occasion suitability, and edge cases—provides actionable data to train agents on the nuances that drive regret. When agents can answer precise fit questions or flag potential mismatches, the likelihood of a return drops, turning the AI layer from a novelty into a profit‑preserving asset.

Operationalizing this insight follows a pragmatic crawl‑walk‑run framework. First, retailers must clean data, quantify bracketing, and capture detailed return reasons beyond generic "didn’t fit" tags. Next, they should link pre‑purchase queries to post‑purchase outcomes, establishing escalation rules when confidence is low. Finally, precision interventions—targeted education, seamless exchanges, or automated guardrails—can be deployed at scale. Companies like True Fit exemplify this approach, leveraging AI to boost keep rates and shrink the $890 billion return burden, positioning confidence‑centric agents as a strategic differentiator in the competitive apparel and footwear market.