AI in childrenswear and basics: forecasting replenishment-driven ranges

KEY TAKEAWAYS Summary by the editors

1. Basics and childrenswear are replenishment-driven: stable styles sell season after season, so AI value comes from forecasting by size and location rather than from predicting trends.
2. Childrenswear adds a specific size problem: children outgrow sizes quickly and parents buy ahead, so demand moves along the size curve over time.
3. Under Article 8 of the GDPR, processing a child's data on the basis of consent for online services requires parental consent below 16, with member states allowed to lower the age to no less than 13.
4. The UK Information Commissioner's Office Children's code sets 15 standards for online services likely to be accessed by children under 18, which matters for any personalisation aimed at young shoppers.
5. For wholesale-supplied basics, the decisive data is re-order and sell-out by store, size and colour; shared with retail partners, it allows automated never-out-of-stock replenishment.

AI in childrenswear and basics is used mainly to forecast demand by size, colour and location, to automate replenishment of never-out-of-stock ranges, and to reduce excess stock in categories with thin margins. The segment differs from trend-led fashion because demand is steady and predictable, so small forecasting gains add up across high volumes, while products for children raise particular data protection and safety responsibilities.

Why does AI matter differently in basics and childrenswear?

Basics such as T-shirts, underwear, socks, leggings and plain knitwear are sold continuously, often in the same styles for years. Childrenswear combines some of this stability (school uniforms, sleepwear, multipacks) with seasonal collections and gifting peaks. In both cases, the business depends on being in stock in the right size at the right place, at a price shoppers find fair.

This changes the AI question. Instead of predicting which new trend will sell, the task is to predict how many of a known item will sell in each size, store and channel, and to trigger replenishment automatically. Value-seeking consumers raise the stakes: McKinsey and The Business of Fashion describe in The State of Fashion 2026 how consumers are redirecting discretionary spending, which puts pressure on price-led categories where availability and value decide the sale.

What are the main AI use cases in childrenswear and basics?

Basics and childrenswear businesses rarely publish detailed AI results, so the table describes common practice rather than named deployments. Each use case should be validated against a brand's own data.

Read also

AI in fast fashion: speed, demand signals and the overproduction problem

How does AI forecast demand for replenishment ranges?

For replenishment ranges, machine learning forecasting works with long, clean histories by item, size and location. It combines past sales with drivers such as seasonality, school calendars, weather, promotions and price changes. The key output is not a single number per style but a forecast per stock-keeping unit and location, translated into order proposals.

A robust replenishment process typically follows these steps:

1. Separate continuous items (basics, multipacks) from seasonal and fashion items.
2. Clean history for stock-outs, because a week with zero stock shows zero sales, not zero demand.
3. Forecast demand by item, size and location, including known peaks such as back to school.
4. Calculate order quantities from forecast, lead time, minimum order quantities and target availability.
5. Let planners review exceptions instead of every line, and record why they override proposals.

What makes childrenswear sizing different?

Children move through sizes quickly, and parents often buy a size up. Demand for a given size is therefore linked to birth rates and the age profile of a store's catchment, and it shifts over the year. Size curves copied from adult categories or from another region can leave stores short of the most needed sizes and overstocked in others. Forecasting by size and location, rather than by style alone, addresses this.

Excess stock is becoming more costly. According to the European Commission, large companies may no longer destroy unsold apparel and footwear from 19 July 2026 under the Ecodesign for Sustainable Products Regulation, and medium-sized companies are expected to comply by 2030. In high-volume basics, even a small structural over-order creates large quantities that must be sold, donated or recycled.

What data and privacy rules apply to childrenswear?

Personalisation needs extra care when children are involved. Article 8 of the GDPR states that where consent is the legal basis for online services offered to a child, processing is lawful only with parental consent below the age of 16, and member states may lower this age to no less than 13. In the UK, the Information Commissioner's Office Children's code sets 15 standards for online services likely to be accessed by children under 18, even if they are not aimed at them.

In practice, most childrenswear customers are adults buying for children, but data about the child, such as age, size or name for personalisation, can still be personal data about a minor. Brands should:

• Collect only the data needed, for example an age band for size advice rather than a date of birth.
• Avoid profiling or targeted advertising directed at children.
• Keep imagery of real children out of AI training datasets unless consent and purpose are explicit.
• Review any generated imagery of children carefully for appropriateness before publishing.

Read also

AI in mid-market and wholesale-led fashion brands: where to start

What risks are specific to this segment, and what should brands do first?

Beyond privacy, the main risks are over-automation (replenishment running unchecked when a style is about to be discontinued), poor handling of peaks such as back to school, and dependence on retail partners' data for wholesale-supplied basics. Product safety information, such as age warnings, must also never be generated without verification.

The first step is data hygiene: correct stock-out periods, unify item and size codes, and identify which items are truly continuous. Then pilot forecasting and automated order proposals on a single basics category with long history, measuring availability, stock levels and planner time. For brands that supply basics to retailers, agree sell-out and stock data sharing with key accounts so that re-orders can follow consumer demand. In this segment, AI rarely looks spectacular, but steady improvements in availability and stock across thousands of items add up.

Frequently asked questions

How is AI used for replenishment in fashion?

AI forecasts demand by item, size and location from sales history and drivers such as seasonality, promotions and weather, then turns the forecast into order proposals. Planners review exceptions rather than every line. It works best for continuous ranges with long, clean histories.

Why is AI well suited to basics and never-out-of-stock ranges?

Basics sell steadily in the same styles for years, which gives models long histories to learn from. Small gains in forecast accuracy repeat across many sizes, stores and seasons. The main data issue is correcting stock-outs so that missing sales are not mistaken for low demand.

Can childrenswear brands personalise with children's data?

Only with care. Under Article 8 of the GDPR, consent-based processing for online services requires parental consent below 16 (member states may set an age between 13 and 16). In the UK, the ICO's Children's code sets 15 standards for services likely to be used by children. Minimising data and avoiding profiling of children are prudent defaults.

How can AI help with childrenswear sizing?

AI can forecast demand by size and store, reflecting how children move through sizes and how age profiles differ by location. It can also support size advice for parents, though this requires careful handling of data about children.

GuideThe complete guide to AI in fashion merchandising and buyingRead the complete guide