Quote Calculation That Survives Volatile Raw Material Prices

When the quote loses money the moment it's sent

Steel, aluminum, copper, plastic granulate, energy, purchased parts: hardly any cost item in manufacturing stays still for more than a few weeks. Yet calculations are often still built on whatever is closest at hand — the material price from the last calculation, the spreadsheet from spring, the master data record in the ERP that nobody has touched in months.

Managing directors and sales leaders in machinery building and supply industries all know the outcome:

• A quote wins — and turns out to be a wash at final costing, because the input material got more expensive in the meantime
• Framework agreements and repeat orders keep running on a year-old price basis
• On complex bills of materials, nobody knows exactly which of the 80 line items is about to tip
• Out of caution, blanket safety margins get added — and those margins cost jobs in competitive bidding

Both directions are expensive: calculate too tight and it eats the margin; calculate too cautious and it eats the order. The real problem isn't the calculation logic — your people have that down. It's the data foundation that nobody can keep current by hand, day after day.

How an AI agent keeps the calculation base current

An AI agent is a digital employee that maintains the price base and prepares calculations — the pricing decision itself stays with sales and management. Here's how it works in practice:

Step 1: Tap price sources that already exist

Current prices are already in the building — just scattered: supplier price lists (PDF, Excel), the most recent purchase orders and order confirmations, daily-price emails from dealers. The agent reads these sources continuously and consolidates them per material and purchased part.

Step 2: Update master data and the calculation base

The agent reconciles detected price changes against the stored calculation prices. Smaller moves within defined limits get updated according to your rules; bigger jumps go up for approval — old and new basis shown side by side. Control stays in-house, but maintenance stops being weekend work.

Step 3: Run quotes against today's prices

For a new inquiry, the agent assembles the calculation on a current basis: bill-of-material items, purchased parts, material usage — valued at the latest prices, with a note on how old each price point is. Your estimator sees at a glance where the basis is fresh and where a supplier price inquiry makes sense — which the agent drafts right along with it.

Step 4: Monitor open quotes and framework prices

The agent checks running quotes and framework prices against price movement: if the margin on an open quote is about to tip because the input material rose, it flags it — while you can still renegotiate or limit the binding period. Price escalation clauses can be backed by data instead of gut feeling.

Step 5: Final costing as a learning loop

After a job is completed, the agent compares planned against actual: calculated versus actual material costs. Systematic deviations — say, a purchased part that consistently comes in pricier than calculated — become visible and feed into the next calculation.

Which systems get connected

HVNH AI's agents work with your existing landscape: ERP, calculation spreadsheets, email inbox, supplier portals, file storage. Your calculation logic — surcharges, overheads, batch-size effects — isn't replaced, just fed with current data. Where no interface exists, the agent works through exports, documents, or the existing program interface.

What realistically comes out of it

A typical result after rollout: price maintenance, which used to cost hours per week or simply didn't happen, now runs in the background. Quotes calculate with this week's prices instead of last quarter's — protecting margin in rising markets and competitiveness in falling ones, because blanket fear surcharges can shrink. Nasty surprises at final costing become rarer, and in price talks with customers and suppliers your team sits at the table with solid numbers.

Worth setting expectations correctly: the agent doesn't set pricing policy. Whether you pass an increase on, absorb it, or hedge it with an escalation clause remains your business decision — it just stops being made blind.

An everyday example

Thursday, 8:30 a.m.: an inquiry for 2,000 assemblies, delivery over six months. The agent assembles the calculation and flags two line items: the last price for the aluminum profile is based on a purchase order six weeks old — the latest dealer price list shows an upward trend; for a purchased part there's a current price list with a 5 percent surcharge that hadn't made it into master data yet. The estimator approves the prepared price inquiry to the aluminum supplier and confirms the master data correction. On Friday the quote goes out — on solid ground, with a binding period matched to the volatility. Time spent in-house: half an hour instead of half a day.

Common objections from practice

"Our calculation is too specific for automation." Your calculation logic stays untouched — the agent delivers current input data and takes over gathering it. The more specific the calculation, the more valuable a clean price base actually becomes.

"Price lists come in twenty different formats for us." That's exactly what AI agents are built for: they read PDF tables, Excel files, and price emails of the most varied structure, and learn the formats of your key suppliers over time. Anything unclear gets flagged for review, never silently adopted.

"Then people will just blindly trust the numbers." That's why the agent shows source and age for every price and clearly separates what's auto-updated from what needs approval. Transparency replaces blind trust — in both directions.

Self-check: is this worth it for your calculation?

• Material costs make up a significant share of your manufacturing costs
• Calculation prices get updated less often than monthly
• Final costings regularly show material deviations
• Price list maintenance keeps getting pushed aside because day-to-day business comes first
• On long-running quotes there's no overview of which line items are about to tip

If three or more of these apply, your calculation base probably holds more margin than any single price negotiation could recover.

The next step

Where your calculation has data gaps today, we figure out in a free intro call: we walk through a real calculation and check how old the price basis of the most important line items actually is. A pilot with one product group follows. More use cases are on our industry page AI in manufacturing.