The Price Rise Letter Customers Actually Read
The move in short
Custom Build — A pipeline triggered when the sales director approves a pricing-round update in the CRM that pulls each affected client's order history, margin profile, tenure, last interaction notes, and the specific product lines being repriced, runs an LLM personalisation to produce a bespoke price-increase letter for each client — acknowledging their specific usage, framing the increase relative to the value delivered, and pre-empting likely objections — and queues each letter for the account owner's review and send approval in a single batched workflow.
The Company
Draken Print & Packaging is a commercial printer based in Gothenburg, about 120 people. They do runs for SME clients across Sweden — labels, folding cartons, branded packaging, short-run marketing materials. A lot of their clients have been with them for years, some over a decade. The relationships are real. Account managers know their clients by name, know their busy seasons, know who calls on a Friday afternoon.
The Pain
Every spring, Erik, Draken's sales director, has to push through a pricing round. Paper costs move, energy costs move, and the numbers have to follow. He knows this. His clients mostly understand it too. The problem is the letters. What goes out is essentially the same block of text to everyone, swapping in a name and a percentage. A client who's been with Draken for nine years, running the same folding carton job every quarter, gets the same letter as someone who started eight months ago. Erik watches a handful of those long-standing clients go quiet after pricing rounds. Sometimes they leave. He suspects they'd stay if someone had just talked to them properly. But writing something real for 60 or 80 clients isn't something he can actually do in the time he has.
The Move
When Erik marks a client as affected by a pricing round in the CRM, a workflow fires. It pulls that client's order history, which product lines are changing, their tenure, their margin profile, and the notes from their last few interactions. It passes all of that to an LLM, which drafts a letter. Not a template with a name dropped in — a letter that mentions how long they've been working together, what they've been buying, what's actually driving the increase, and where relevant, what Draken has done to absorb costs before passing them on. It also addresses the objections that tend to come up, because Erik knows what those are, and they can be written in.
Each draft lands in a review queue. The account owner reads it, tweaks it if they want to, and approves it. The whole batch can go out in a morning instead of across three weeks of Erik chasing people to write letters they don't have time to write well.
The blind spot
Most people assume you'd need to do this one client at a time to make it personal. The instinct is: personalisation means slow. But the personalisation here is happening in the data pull, not in someone sitting down to draft. The LLM is just assembling what's already in the system. Once the workflow exists, the time per letter is mostly the account owner's review, not the writing.
The pattern
The same setup works in a few other ordinary situations:
- A professional services firm sending renewal proposals to retained clients at different rates
- A logistics company notifying customers of route or rate changes that affect them differently depending on their contract
- A software reseller communicating a vendor price increase across a customer base where usage and contract length vary a lot