AI Pricing Strategies for Retail Tariffs in 2025

Retail tariffs don’t feel like a “pricing” problem when you’re the one staring at landed cost spreadsheets at 11 p.m. They feel like a survival problem.

A December 2025 survey found 54% of online retailers are impacted by tariffs, and 39% have already raised prices with more increases planned. Nearly 1 in 5 cut the number of products they sell, and about 20% raised prices by more than 10%. The pattern is clear: when costs jump and margins are thin, retailers default to blunt instruments—price hikes, SKU cuts, and fewer perks like free shipping.

Here’s the thing: those moves are understandable, but they’re also often sloppy. Tariffs create a decision-quality crisis as much as a cost crisis. If you can’t see which customers will tolerate which price moves, or which SKUs actually earn their shelf space across channels, you end up cutting the wrong items and discounting the wrong ones.

This post is part of our AI in Retail and E-Commerce series, and I’m going to take a stance: tariffs don’t force bad customer experiences—bad measurement does. AI can’t erase tariffs, but it can help retailers adjust pricing, product assortment, and omnichannel operations with less collateral damage.

Why tariffs push retailers into “bad” decisions

Tariffs raise costs, but the real damage comes from how fast they compress your options. When margins are already squeezed by shipping, labor, and paid media, you don’t have months to test and learn. You make moves quickly—or you run out of cash.

Most retailers respond with three common plays:

• Raise prices (often broadly, often too much, often all at once)
• Shift suppliers (valid, but operationally messy and slow to stabilize)
• Reduce product selection (fast relief, long-term brand risk)

The hidden problem: each of these actions changes demand. Raising prices changes conversion. Cutting SKUs changes basket size. Switching suppliers changes reviews, return rates, and repeat purchases. If you don’t model those second-order effects, you’re basically guessing.

A more disciplined approach is to treat tariffs as a trigger to tighten your feedback loops:

• Which price changes reduce profit because volume collapses?
• Which SKUs look unprofitable until you account for cross-sell?
• Which customers actually care about free shipping, and which care about speed?

That’s the gap AI fills well—not because it’s magic, but because it handles complex trade-offs faster than manual analysis.

AI pricing optimization: raise prices without losing the wrong customers

Direct answer: AI helps retailers raise prices more safely by estimating price sensitivity at the SKU and segment level, then recommending smaller, targeted price moves instead of broad increases.

Stop treating price elasticity like a single number

Many teams still use one “price sensitivity” assumption for a whole category. That’s how you end up overpricing your traffic drivers and underpricing your premium niche.

A practical AI pricing setup groups demand into layers, for example:

• Must-have essentials (low elasticity, but high trust sensitivity)
• Comparable commodities (high elasticity; shoppers will price-check)
• Differentiated products (elasticity depends on brand/story/reviews)
• Add-ons (often low absolute revenue, high basket impact)

With that, you can make tariff-driven increases more precise:

• Increase prices where customers won’t churn
• Hold or even reduce prices on “storefront SKUs” that anchor trust
• Shift margin recovery to bundles and add-ons where it’s less visible

What “good” looks like: targeted increases and guardrails

If you’re going to raise prices (and many retailers will), set guardrails so you don’t accidentally torch acquisition.

Use AI to monitor and react to:

• Conversion rate by segment (new vs returning, price-sensitive cohorts)
• Profit per session (not just gross margin)
• Cart abandonment reasons (especially shipping + total cost)
• Competitor price gaps (for your top traffic SKUs)

A stance I’ll defend: price increases should be staged, not dumped. Even when your cost jump is immediate, your customer learning curve isn’t. Gradual moves paired with clearer value cues (delivery promises, warranties, easy returns) beat sudden shocks.

Personalization that doesn’t feel creepy

AI-driven personalization can offset price pain if it’s used to improve relevance rather than manipulate.

Examples that help in tariff seasons:

• Recommend lower total-cost alternatives (not just cheaper items—cheaper delivered)
• Offer bundle suggestions that preserve perceived value
• Highlight in-stock local inventory to reduce shipping exposure

Customers don’t mind being guided. They mind feeling trapped.

Assortment planning: don’t cut SKUs until you know what they’re really worth

Direct answer: AI reduces “panic SKU rationalization” by measuring each product’s total contribution—including returns, customer lifetime value impact, and basket attachment—not just unit margin.

The survey finding that 19% of retailers have cut the number of products they sell is believable because SKU reduction is fast. But it’s also easy to do badly.

The classic mistake: killing the items that bring people in

Some SKUs are loss leaders in isolation but profitable as traffic magnets or bundle anchors. When tariffs hit, teams often cut these because their margins look worst—right when acquisition is already expensive.

A more complete SKU score should include:

• Contribution margin after shipping and pick/pack
• Return rate and return cost (especially for apparel/footwear)
• Attach rate (what else gets bought with it)
• Substitution behavior (what customers buy if it’s unavailable)
• Review impact (removing a top-rated item can hurt category trust)

AI models can estimate these relationships using order history, clickstream behavior, and inventory events (stockouts, substitutions, backorders).

Three AI-driven ways to avoid unnecessary product cuts

1. Predict substitution before you delist

   • If customers will switch to a higher-margin substitute, delisting may be safe.
   • If they’ll leave the site, you’re about to pay the tariff twice—once in cost, once in churn.

2. Use “good-better-best” ladders

   • Keep a clear value option, a mainstream option, and a premium option.
   • AI can help set the price spacing so the “better” option becomes the default.

3. Rationalize variants, not hero SKUs

   • Cut slow-moving colors/sizes/pack sizes first.
   • Keep the product that earns search demand and reviews.

If you need margin relief quickly, variant trimming usually damages your brand less than deleting entire product families.

Omnichannel inventory: tariffs punish stockouts more than you think

Direct answer: AI improves omnichannel resilience by forecasting demand more accurately under price changes and allocating inventory to the channels that maximize profit and customer satisfaction.

Tariffs don’t just raise costs—they increase the cost of being wrong.

When your landed cost rises, you can’t afford:

• Overstock (capital tied up in higher-cost inventory)
• Understock (lost sales you can’t cheaply reacquire)
• Misallocated inventory (sitting in the wrong location/channel)

Demand forecasting under price changes

A basic forecast assumes “tomorrow looks like yesterday.” Tariff-driven pricing breaks that.

AI demand models can incorporate:

• Price changes and promo calendars
• Competitor price movements
• Seasonality (yes, even in late December, when gift cards and returns reshape demand)
• Channel effects (online vs store vs marketplace)

One practical win: simulate outcomes before you change prices. If a 7% increase drops unit demand 12% for a subset of SKUs, you’ll see the fulfillment impact before you create a stock problem.

Smarter allocation across channels

Retailers often treat channels as separate P&Ls. Customers don’t.

AI allocation can support decisions like:

• Ship-from-store vs ship-from-warehouse based on margin and delivery promises
• Reserving scarce inventory for higher-LTV segments or higher-margin channels
• Rebalancing inventory ahead of predictable return waves (post-holiday is brutal)

And yes, there’s a customer experience angle: fewer stockouts and fewer cancellations do more for loyalty than almost any coupon.

Customer experience during price hikes: be honest, then be helpful

Direct answer: AI supports customer experience by identifying friction points caused by tariffs (shipping thresholds, delivery speed sensitivity, price shock) and tailoring offers and messaging to reduce churn.

Retailers often fear explaining price changes. I think silence is worse. Customers notice. They’ll fill the gap with their own narrative, and it won’t be kind.

What to communicate (and what not to)

If tariffs push you to adjust pricing or reduce “free shipping,” focus on clarity:

• Set expectations on shipping thresholds and delivery times
• Provide alternatives (pickup, slower shipping, bundles)
• Avoid long policy pages nobody reads

Where AI helps is choosing who needs reassurance:

• New visitors might need visible price-matching signals or stronger value proof.
• Returning customers might respond better to loyalty perks that preserve their usual experience.

A simple retention play that works

If you’re raising prices, don’t treat retention as “send more emails.” Treat it as protecting the relationship.

AI can flag customers at risk (reduced browse frequency, lower basket size, more comparison clicks) and trigger practical interventions:

• Early access to restocks
• Personalized bundles that hold total order value steady
• Targeted free shipping on the customers who will actually churn without it

Blanket discounts are expensive. Targeted reassurance is cheaper.

A 30-day AI action plan for tariff pressure

Direct answer: Start with pricing + assortment analytics, then expand to inventory allocation once measurement is stable.

If you’re trying to respond quickly, here’s a tight sequence I’ve found realistic for teams that aren’t starting from scratch.

Days 1–10: Get decision-grade cost and margin visibility

• Build a single view of landed cost by SKU (tariff included)
• Calculate true contribution margin (shipping, pick/pack, returns)
• Identify the top 50 SKUs by traffic and by profit (they’re not always the same)

Days 11–20: Pilot AI pricing on a controlled slice

• Choose 1–2 categories with stable demand and clear substitutes
• Set guardrails: max price move, max conversion drop, minimum profit/session
• Run staged increases and track segment-level response

Days 21–30: Assortment and omnichannel tuning

• Score SKUs using attach rate + substitution risk
• Trim variants first; pause reorders on low-value tails
• Reallocate inventory to reduce cancellations and late deliveries

The goal in month one isn’t perfection. It’s avoiding the worst outcomes: pricing yourself out of demand, or cutting the products that keep customers coming back.

Where this fits in the AI in Retail and E-Commerce series

Tariffs are forcing retailers to confront a hard truth: your pricing, assortment, and fulfillment decisions are only as good as your data and your ability to act on it quickly. That’s exactly why AI keeps showing up in modern retail stacks—customer behavior analysis, personalized recommendations, pricing optimization, and omnichannel execution are no longer “nice to have” when costs spike.

If your 2026 plan is “raise prices and hope,” you’re leaving money—and loyalty—on the table. A smarter plan is to use AI to be more selective: raise prices where you can, protect the SKUs that earn trust, and keep the omnichannel experience reliable even when unit economics get ugly.

What would change in your business if you could answer this with confidence: which 20% of price moves protect 80% of your margin—without driving away the customers you can’t afford to lose?