How is AI used in fashion sourcing for cost modelling and supplier negotiations?
AI can break down garment costs, compare quotes and prepare negotiations. What works, what consultancies claim, and why trust and supplier relationships still decide outcomes.
KEY TAKEAWAYS Summary by the editors
- In fashion sourcing, AI is mainly used to build should-cost models, normalise and compare supplier quotes, analyse spend and prepare negotiations.
- Walmart used AI-powered negotiation software with a text interface to negotiate with tail-end suppliers, and Harvard Business Review reported that it closed agreements with 68 percent of suppliers approached.
- BCG's 2025 procurement study reports savings of up to 15 percent in some categories and tender drafting around 50 percent faster in a client case, but these are consultancy estimates rather than independent results.
- Should-cost models are only as good as their inputs: fabric consumption, material prices, labour minutes, overheads and duties must be current and specific to the product.
- Automated negotiation fits standardised, lower-value purchases best; strategic garment suppliers depend on long-term relationships, capacity planning and compliance, which AI supports but does not replace.
AI is used in fashion sourcing to estimate what a garment should cost, compare supplier quotes on a like-for-like basis and prepare buyers for negotiations with data on materials, labour and market prices. Automated negotiation agents already exist for standardised purchases, as Walmart's example shows. For strategic garment suppliers, AI is best seen as preparation and analysis support, not as a negotiator.
What is should-cost modelling for garments?
A should-cost model breaks a product's price into its components: fabric and trims (consumption multiplied by price), cut, make and trim labour (minutes multiplied by labour rate), washing or finishing, overheads, margin, packaging, freight and duties. Comparing the model with a supplier's quote shows where the quote looks high or low and gives the buyer a fact base for discussion.
Building these models manually for every style takes time, so many companies use them only for core products. AI can extend coverage by reading tech packs and bills of materials, estimating fabric consumption from patterns or similar styles, and pulling current material prices and labour rates into the calculation.
| Application | What AI does | Key inputs | Main risk |
|---|---|---|---|
| Should-cost modelling | Estimates component costs per style | Tech packs, BOMs, material prices, labour minutes and rates | Outdated or generic inputs |
| Quote normalisation | Converts supplier quotes in different formats into comparable cost breakdowns | Quotes, incoterms, currencies | Misread conditions or currencies |
| Spend analytics | Classifies purchases and shows price variance across suppliers and seasons | Purchase orders, invoices | Poor product categorisation |
| Negotiation preparation | Summarises supplier history, benchmarks and talking points | Performance data, contracts, market data | Over-confident recommendations |
| Automated negotiation | Negotiates standard terms with many small suppliers | Pre-approved trade-offs and limits | Damaged relationships if misused |
Can AI negotiate with suppliers?
It can, within limits. Harvard Business Review described how Walmart, which has more than 100,000 suppliers, used AI-powered negotiation software with a text-based interface to reach tail-end suppliers whose standard agreements were often never negotiated because buyers lacked the time. The chatbot was negotiating and closing agreements with 68 percent of suppliers approached, with each side gaining something it valued. The authors recommended starting with indirect spend and pre-approved suppliers and deciding acceptable trade-offs in advance.
Fashion's strategic suppliers are different. Garment factories commit capacity, invest in skills and carry compliance obligations, and prices are linked to lead times, minimums and quality. Automated haggling over a few cents per unit could undermine relationships that matter for allocation in tight seasons and for social and environmental standards.
What benefits do consultancies report?
Consultancy studies are optimistic. BCG's February 2025 perspective on procurement with AI reports immediate savings of up to 15 percent depending on category, mainly in indirect categories such as IT and marketing, and describes a client case where tender drafting became about 50 percent faster. Roland Berger argued in 2023 that generative AI could support strategy development, tender management, negotiation preparation and contract management, and that chief procurement officers could achieve better results with 50 to 75 percent of current resources.
These figures should be read as indications from consultancies' own client work, often stated as 'up to', and not as independent benchmarks for fashion. Direct material sourcing in apparel, with its complex cost structures and supplier dependencies, is likely to see smaller and slower effects than indirect spend.
BCG also attributes transformation success mainly to people and only partly to technology and algorithms, a reminder that tools alone do not change outcomes. Buyers need to understand how a model reaches its numbers, merchandisers need to trust the cost targets, and suppliers need to see that the process is consistent. Where these conditions are missing, sophisticated models tend to end up unused or overridden.
How does AI help compare supplier quotes?
Quote comparison is often the quickest win. Suppliers send prices in different templates, currencies and delivery terms, sometimes with fabric and making costs combined, sometimes split. A language model can read these documents, map each element to a standard cost breakdown, convert currencies and incoterms using rules defined by the sourcing team, and highlight where a quote is incomplete. The buyer then sees comparable numbers side by side and can focus questions on the elements that differ most, such as fabric price per metre or making minutes.
What data does AI-assisted sourcing need?
- Structured tech packs and bills of materials with consumption figures.
- Current prices for main fabrics, yarns and trims, with dates and sources.
- Standard minute values or labour times per operation, and labour rates by country.
- Historical quotes, purchase prices and supplier performance data.
- Freight, duty and currency assumptions, kept up to date as conditions change.
In 2026, the last point is critical. McKinsey's State of Fashion 2026 names tariffs as the top-cited hurdle among executives and expects brands to respond with pricing, sourcing shifts and efficiency gains. A should-cost model with outdated duty assumptions can be more misleading than no model at all.
How should a sourcing team introduce AI?
- Standardise cost breakdown templates so supplier quotes share a common structure.
- Use AI to normalise and compare quotes for one product category.
- Build should-cost models for core styles and compare them with actual negotiated prices.
- Introduce negotiation preparation briefs that summarise data for buyers.
- Consider automated negotiation only for standardised, low-risk purchases such as packaging or indirect services.
- Review results with finance and with suppliers to make sure the process is seen as fair.
What are the risks?
The main risks are inaccurate inputs producing confident but wrong cost targets, opaque models that buyers cannot explain to suppliers, and pressure on supplier margins that conflicts with responsible purchasing commitments. Cost models that drive prices below what is needed to pay fair wages and meet standards create compliance risks elsewhere. Good practice is to keep models transparent, share cost logic with strategic suppliers where appropriate and include labour cost assumptions that reflect real conditions.
Frequently asked questions
What is should-cost analysis in fashion?
It is an estimate of what a garment should cost based on its materials, labour time, overheads, freight and duties. Buyers compare it with supplier quotes to identify gaps and prepare negotiations.
Did Walmart use AI to negotiate with suppliers?
Yes. Harvard Business Review reported that Walmart used AI-powered negotiation software with a text-based interface for tail-end suppliers and closed agreements with 68 percent of suppliers approached.
How much can AI save in procurement?
BCG reports savings of up to 15 percent in some categories, mainly indirect spend, based on its client work. These are consultancy estimates, and results for direct materials in fashion are likely to vary widely.
Should fashion brands automate negotiations with garment factories?
For strategic garment suppliers, automated negotiation carries relationship and compliance risks. AI is better used there to prepare buyers with data, while automation suits standardised, lower-value purchases.
One edition every weekday morning. Read in five minutes. Free for industry professionals.
