Office & managementVerified Real deployment at a named or precisely described company, with a measured result from a primary source.

AI price engine predicts which price each customer will accept, so price increases stick without losing volume (Hilti, Fabory, a building products manufacturer)

Added 2026-11-16 · Last updated 2026-11-16

Manufacturers in this case: Hilti · all manufacturers using AI

Role

Sales, pricing and finance (list prices, discounts and annual price increases) · Typical users: pricing manager, sales director, product manager, CFO or controller, general manager of a manufacturer selling to businesses

The problem

Many manufacturers set a list price and then let each salesperson decide the discount, or raise all prices by the same percentage when costs go up. Prices drift apart for similar customers, discounts are given too early, and across-the-board increases do not stick. At Hilti, new account managers began every sales call with questions such as what price to give and whether it was fair compared to similar customers, and pricing had become transactional and hard to explain. At a US building products manufacturer (unnamed in the Zilliant case study), sales quoted too far below the price book in anticipation of customer pushback, and almost half of monthly transactions deviated enough to require lengthy and costly internal price reviews; cost- and margin-based increases did not stick. At Fabory, a European supplier of fasteners, rising shipping and raw material costs after the pandemic and new EU anti-dumping duties on stainless steel imports forced price increases across the whole portfolio in 2021.

How AI helps

The model learns from the company's own quotes, orders and customer data how likely a customer is to buy at a given price, then recommends a price per product and customer segment that people can review before it goes out. At Hilti, quotes and the orders that followed were treated as yes or no decisions at specific price points; a machine learning model built by an internal data science team with external partners takes product attributes, customer characteristics and deal features, estimates the probability of purchase for each deal at different prices, and adds these up to simulate volume and revenue for a product range. It is used to reprice product categories in a market and to simulate volume when planning annual price changes, behind a price engine that sets prices by customer trade and five customer potential classes. Outputs were checked against real transactions and in pilots before wider rollout. At the building products manufacturer, Zilliant Price IQ applied price optimization to customer, product and order data, taking customer price sensitivity and market prices into account; the pricing team adjusted each product line in Zilliant's strategy interface before publishing a new price book, about three months after starting. At Fabory, the consultancy Rewire built an AI pricing engine that calculates a price for each product and models how changes in input costs and other external factors affect demand and margin.

Step by step

  1. Export two or three years of quotes and orders with customer, product, quantity, list price, net price, discount and whether the quote was won or lost.
  2. Measure the spread today: for similar customers and products, how far apart are net prices, and how many deals go below the price book or need an exception?
  3. Agree on simple pricing principles with sales and management (for example, similar customers pay similar prices) and on customer segments the sales team understands.
  4. Build or buy a model that estimates win probability against price for each segment, and test it on past deals it has not seen.
  5. Pilot on a few product lines and keep a comparison group priced the old way, so the margin and volume effect can be measured.
  6. Keep people in charge: product or pricing managers approve the recommended prices, sales can request exceptions, and exceptions are tracked.

Tools

  • In-house machine learning purchase probability model with an automated price engine: used at Hilti
  • Zilliant Price IQ price optimization: used at a US building products manufacturer
  • AI pricing engine built by Rewire (data and AI consultancy): used at Fabory
  • Other tools to evaluate: price optimization and CPQ software from pricing suppliers, the pricing module of your ERP or CRM, Python or your BI tool on your own quote and order history
  • To start without new software: a quote and order history export and any LLM (ChatGPT, Claude, Copilot, Gemini) with the prompt below

In the AI tools index: Zilliant Price IQ · all AI tools used by manufacturers

Result

Hilti (Copperberg article on a keynote by Hilti's head of its Fair Pricing Competence Center, 2026): more than half of global revenue now runs through the automated price engine, about 40% through individual price agreements for key accounts and around 10% through one-off overrides; no margin figure was published. Building products manufacturer (Zilliant case study, the software vendor): a sustained 2.3% margin increase on the product lines priced with Price IQ while keeping historical volumes, against 1.0% for lines of the same brand raised with the old across-the-board method: 130 basis points better, or USD 3.4 million a year. Fabory (Rewire case study, the consultancy): a pilot across the full portfolio in Belgium and the Netherlands raised prices without eroding volumes, with an EUR 8 million margin uplift, a 6% revenue increase and 33,000 contracts changed.

As reported by the source; most figures are vendor or self-reported and not independently verified.

Human review: cautions

  • The building products and Fabory figures come from the vendor (Zilliant) and the consultancy (Rewire) that sold the work; the building products manufacturer is not named. Hilti published no margin figure.
  • Fabory is a fastener supplier and distributor rather than a producer; the method applies to manufacturers with many SKUs and customers, but your result will differ.
  • A pricing model needs enough won and lost quotes; with few customers or one-off products, start with price consistency analysis rather than prediction.
  • Executive backing and sales adoption mattered as much as the model in these cases: Zilliant credits the vice president of sales, and Hilti rolled out simpler analytics first.
  • Check competition law: prices must come from your own data and decisions, never from sharing prices with competitors.

Try it yourself

You are a pricing analyst in a manufacturing company that sells to businesses. I want to find where our prices are inconsistent and where a price increase is most likely to stick.

I will paste: (1) a quote and order history export [quote date, customer, customer segment or size, region, product, product family, quantity, list price, net price, discount %, won or lost]; (2) our current cost increase per product family, if any.

Do this:
1. For each product family and customer segment, show the median net price, the spread (lowest to highest) and the share of deals below a discount threshold I will give you.
2. List the customers who pay much less than similar customers for the same products, with the size of the gap.
3. Where we have enough won and lost quotes, describe how the win rate changes with the discount level, and say where the data is too thin to conclude.
4. Propose a price increase per product family and segment instead of one flat percentage, and estimate the margin effect if volumes stay the same.
5. Draft a pilot plan with a comparison group priced the old way, and the measures to track (margin, volume, win rate, exceptions).

Do not suggest prices for a customer based on anything other than the data I paste. Use only the data I paste in; if something is missing, write 'not found' or list it as a question instead of guessing. Do not send, order, post or change anything in any system: produce drafts for my review and approval.

Want more like this? Join the newsletter

Related cases

All 15 Office & management use cases →

Sources

Weekly: 1–2 practical AI use cases for your plant

Newsletter in

Free, one email a week. By subscribing you agree to receive the weekly newsletter (and its sponsor message) from AI Agents for Manufacturing. Every email includes an unsubscribe link, and you can unsubscribe anytime. See our Privacy Policy.