How to Use AI to Handle a Supplier Price Increase

When a supplier raises prices, the first useful move is not negotiation and not a price rise of your own. It is working out which of your products the increase actually damages. A 9% increase on an input that is 40% of one product's cost and 4% of another's produces two completely different problems, and most small businesses respond to both the same way because nobody did the arithmetic.

AI is genuinely good at this specific job, because it is arithmetic across a product list that you have been avoiding since the letter arrived.

Start with the exposure, not the percentage

The number in the letter is not your number. Your number is how much of each product's cost that input represents.

Put your product list into a spreadsheet with three columns: selling price, current unit cost of the affected input, and how many units of that input each product uses. Then ask for the margin impact:

For each product below, calculate:
1. Current gross margin (price minus total cost, as a % of price)
2. New unit cost if [INPUT] rises by [X]%
3. New gross margin at the same selling price
4. Margin points lost
5. The price rise needed to hold the current margin %

Sort by margin points lost, largest first.
Show your working for the first row so I can check it.

[paste table]

That "show your working for the first row" line matters. Check it by hand. If the first row is right, the method is right, and you can trust the sort order. If it is wrong, you have caught it in ten seconds rather than after you have repriced your catalogue.

The output almost always surprises people. The products that hurt are rarely the ones you expected, because exposure depends on the input's share of cost, not on how much of the input you buy in total.

The three responses, and the numbers that choose between them

Every product lands in one of three buckets. The point of the exposure table is that it assigns them for you.

Absorbing is a real choice, not a failure to act. A loss-leading item that brings people in can carry a couple of margin points if the basket around it is healthy. What you must not do is absorb by default across everything, which is how a business discovers eighteen months later that it is working the same hours for two thirds of the profit.

Passing through is the default for anything standalone. The question is not whether, it is how much and how you say it. Holding your margin percentage usually requires a slightly larger rise than people expect, because you are marking up the new cost, not adding the increase.

Re-speccing means changing what you buy: a different grade, a different supplier, a different quantity break, or reformulating so you use less. This is where the effort belongs when one input dominates a product's cost, and it is the response most small businesses reach for last because it takes real work.

Use AI for the parts that are research, not judgement

Three jobs it does well here.

Checking whether the increase is market-wide or specific to you. If your supplier's costs genuinely rose, so did their competitors'. If they did not, you have leverage. Producer Price Index series by commodity, published by the Bureau of Labor Statistics and browsable through the St. Louis Fed's FRED database, tell you what actually happened to input prices in that category. Ask for the relevant series and read the trend yourself rather than accepting a summary. A supplier citing "market conditions" for a 12% rise in a category where the index moved 3% has given you your opening line.

Drafting the questions for the supplier. Not the negotiation, the diligence. What drove the increase, which components of the cost moved, whether volume commitments change the number, whether a longer contract fixes it, what the notice period is, and whether there is a grandfathering window on existing orders. A model will generate a thorough version of this list faster than you will remember all of it under irritation.

Modelling the volume trade. Suppliers frequently offer to hold the old price against a larger commitment. Whether that is good depends on your carrying cost and how fast the stock moves, which is a calculation, not an instinct. Give it your turnover rate and ask for the break-even holding period.

What it should not do is decide. The absorb-or-pass-through call depends on things not in your spreadsheet: which customers are price-sensitive, which competitor is likely to hold, and what your reputation can take. Those are yours.

Then handle the customer side

If you are passing the increase through, the notice is its own piece of work, and doing it badly costs more than the increase did. Writing a price increase letter with AI covers the grandfathering arithmetic and the stock phrases to cut.

Two rules worth stating here. Give the date and the number in the first two sentences. And do not blame your supplier: customers do not care about your cost structure, and it reads as an excuse rather than a reason.

The pattern worth keeping

Do this once properly and you have built a reusable model. The spreadsheet, the exposure calculation, and the three-bucket rule work for the next increase and the one after. Keep the file.

The same exposure thinking applies well beyond suppliers. It is how you decide which products to discontinue, and it pairs with forecasting demand to tell you whether a re-spec is worth the switching cost. If the increase has you looking at your own overheads, auditing your tool spend is usually the fastest place to find offsetting money.

And before you accept the new price at all, it is worth one conversation. Negotiating better supplier rates with AI covers the prep, which is where negotiations are actually won. More on running the rest of the business this way in the AI for small business guide.

FAQ

How do I work out what a supplier price increase really costs me?

Calculate the increase as a share of each product's total unit cost, not as a headline percentage. The same 9% rise can cost one product five margin points and another half a point.

Should I pass a supplier increase on to customers?

Usually yes for standalone products losing three or more margin points. Absorb only where the product pulls through other sales and the loss is small, and be able to name what it pulls through.

How much should I raise prices to cover an increase?

Slightly more than the cost increase itself if you want to hold your margin percentage, because you are marking up the new higher cost rather than adding the difference to your old price.

Can I check whether the increase is justified?

Partly. Producer Price Index data by commodity shows what actually happened to input prices in that category. A rise far above the index is worth questioning directly.

Is it worth committing to higher volume to hold the old price?

Only if the carrying cost of the extra stock over your real turnover period is less than the saving. That is a calculation, and it needs your actual stock turn rate rather than an estimate.