Customer story · Loader cranes · DenmarkHow HMF Cranes lets a conversation do the first part of quoting a crane
HMF's dealers know how to configure a crane. The harder part is translating a customer's job into what to propose. HMF added an AI sales assistant from Sailsrep on top of its existing CPQ, so a quote can start with a plain-language conversation and end as a valid, priced proposal. From the first project meeting to a live end-to-end flow took six weeks, in internal testing before the dealer rollout.
Magnus Fasth Co-founder, Sailsrep
Worked hands-on with HMF's team on this project · Updated August 2026
- Built in Denmark since 1952
- 3 to 95 tonne-metres
- Dealers in 30+ countries
Dealers quote for themselves
HMF's distributors used to spend a lot of time making quotes for the dealers. With the assistant, dealers create the quotes themselves.
A senior salesperson in every conversation
A dealer is not an expert in HMF cranes. The assistant works the way a senior HMF salesperson would, bringing that product knowledge to dealer conversations across 30+ countries.
Never an impossible crane
HMF's product rules are held in a constraint solver, so every configuration the conversation reaches is one HMF can actually build.
From conversation to CPQ
A chat produces a dealer quote that flows straight into HMF's CPQ, and the customer receives an interactive proposal page with configuration and price.
A light on the sales conversation
Which questions actually get asked when a crane is sold is of great value to know. The assistant puts a light on that conversation, and the analytics, once this is in use, are the interesting next step.
A crane quote starts with a job, not a spec sheet
HMF has built truck-mounted loader cranes in the Aarhus area of Denmark since 1952. The lineup spans lifting capacities from 3 to 95 tonne-metres, sold through dealers in more than 30 countries.
A customer rarely arrives asking for a specific model. They arrive with a job, what they lift, where, how often, what truck it has to sit on, and at most a basic requirement like how much the crane needs to lift. Turning that into the right crane takes product knowledge that traditionally lives in the heads of experienced salespeople. HMF sells through dealers in 30+ countries, and that knowledge has to reach every one of those conversations. Most dealers sell more than one brand, and Lars Hansen names it as one of HMF's hopes: give dealers a better tool than the competition does, and the HMF crane becomes the easier one to sell.
“Our dealers spend a lot of time making quotes, and a dealer is not an expert in HMF cranes. That is why they need this senior salesperson to help.”
The assistant sits on top of what HMF already runs
HMF added the Sailsrep sales assistant as a conversational layer on top of the systems they already run: their existing CPQ, Microsoft Dynamics 365, and product data in their PIM. Nothing was replaced.
The PIM is no side system. HMF has run it for six years, and it feeds the public website, which keeps a standing pressure on the data to be correct. In Lars's words: "It was obvious to use it for other things."
Sailsrep sales assistant
HMF's existing CPQ
Completed conversations create dealer quotes here
Microsoft Dynamics 365
The CRM HMF already runs
PIM system
Six years of product data, synced into the assistant
Nothing replaced: the assistant connects to the CPQ, CRM and product data already in place.
It works the way a senior HMF salesperson would
- 01
Understand the use case
What the customer lifts, where, how often, and what truck it has to sit on.
- 02
Work out the numbers
The reach and lifting capacity the job actually needs.
- 03
Propose the right model
A crane HMF can build, configured and priced.
A large language model handles the conversation; a constraint solver enforces HMF's product rules, so the dialogue can never end in a crane that can't be built. A completed conversation creates a dealer quote in the CPQ, and the customer receives an interactive proposal page with configuration and price.
“The data comes from our own systems, and the constraints control what can be built. That safe frame has been a big thing in this project. Without it, we would not have proceeded.”
Six weeks to a working flow
The project started with a two-day, hands-on workshop, and the teams have run it together since. Within six weeks of the very first project meeting, the end-to-end flow, conversation to dealer quote to priced proposal page, was live in internal testing. HMF has been through full-scale CPQ implementations, and the time to value here is what stood out: the assistant started proving itself while a traditional project would still have been in requirements gathering.
Fast time to value
Six weeks from the very first project meeting to a live end-to-end flow, not a year-long CPQ project.
Nothing replaced
Sailsrep runs as a layer on top of HMF's existing systems, with six years of product data synced from their PIM.
Capability stays in-house
HMF and Sailsrep ran the project together. Along the way HMF's team learned the tools well enough to make adjustments themselves.
“The most surprising part was the workshop, when we sat and actually tried to use it. And then how flexible the system is. You can do so many things, and you see the immediate result.”
Where the project stands
HMF is now in the testing phase. When we sat down with Lars in August to update this story, internal flow tests with the export department's sales reps had produced one complaint so far: the assistant may talk too much. HMF does not know yet whether that is a flaw or a feature. A dealer who does not know every crane detail may want exactly that level of explanation, so the dealer pilot will decide.
Lars had just put together the sales packages himself, the kind of adjustment HMF's team now makes on its own. Next comes rollout to the dealer network.
- Done
Two-day workshop
- Done
End-to-end flow live
- Happening now
Internal testing with export sales reps
“Headquarters does not know what questions are actually asked when a crane is sold. This puts a light on that conversation. The analytics part, once this is in use, is going to be an interesting next step.”
Frequently asked questions.
Did HMF replace its existing CPQ?
No. Sailsrep runs as a conversational layer on top of the CPQ, CRM and PIM HMF already runs. Completed conversations create dealer quotes directly in the CPQ.
How can an AI be trusted to quote a configurable product?
The language model never decides what is valid. A constraint solver holding HMF's product rules does. The AI contributes the conversation and the reasoning; the solver guarantees every proposal is a crane HMF can build. For HMF, that guarantee was the condition for doing the project at all.
How long did setup take?
A two-day workshop got HMF's team hands-on with the tools, and within six weeks of the very first project meeting the end-to-end flow was live in internal testing. Sailsrep aims at about 40 hours of effort to a working product; HMF's integrations added a bit on top of that.
Does this require ripping out existing systems?
No. This project is the "improve what you already have" path: the AI layer connects to the CPQ, CRM and PIM already in place.
Was it worth doing on top of an existing CPQ?
When we asked Lars Hansen whether a manufacturer of HMF's size, already running a CPQ, should do this, the answer was an immediate yes. The assistant does not replace the CPQ investment, it starts more of the quotes that end up there.
Already have a CPQ? Give it a front door.
If your dealers or customers could start a quote by simply describing the job, more quotes would get started. It can be live in weeks. Book a meeting and see the assistant on your own product.
Book a meetingSWIFT Lifts: 10x more quote requests with an AI sales assistant
How SWIFT replaced its web configurator with a conversation and grew quote requests from about 2 per week to about 20.
Read the storyError-free quotes in 35 markets
HMF's CPQ implementation, told by cpq.se.
Read it on cpq.se, opens in a new tab