Your online configurator is a salesperson, not an order form
Most online configurators collect choices into a form. B2B buyers want a priced answer. What changes when the configurator sells, and how to measure it.
Most manufacturers I talk to already have something called a configurator on their website. Pick a model, tick some boxes, leave your email, and somebody gets back to you in a few days. That is not a configurator. It is an order form with nicer buttons, and buyers can tell the difference.
Why are B2B buyers configuring before they call?
Because they have stopped wanting to call first. Gartner's 2026 sales survey of 646 B2B buyers found that 67% prefer a rep-free buying experience, up from 61% in the previous year's survey. 45% had used AI at some point in a recent purchase. McKinsey's 2024 B2B Pulse research puts the number of channels a buyer uses in a purchase at around ten, and has found roughly one in five decision makers comfortable spending between $500,000 and $5 million in a single remote or self-service transaction.
These are not consumer habits leaking into small B2B transactions. This is how capital equipment is starting to be bought. Automotive showed where it leads: car buyers do their research online and arrive at the dealer with a configured car in mind. The configurator did not replace the dealer. It changed what the dealer is for: validation and closing, not information.
What is wrong with a classic web configurator?
A classic configurator collects choices, but it does not sell. The buyer gets a form when they really want a knowledgeable salesperson. It cannot answer "which one do I need for this?", it cannot explain why an option just disappeared, and it ends by asking for an email address instead of giving an answer. Every hard question goes back into a queue for a human, usually the next working day.
The buyer, meanwhile, is on a phone at a quarter to ten in the evening, asking the questions they do not yet want to ask a person. Whoever answers them first, correctly and with a price, gets the conversation.
What happened when SWIFT replaced its configurator?
SWIFT Home Lifts had exactly that kind of classic web configurator. They replaced it with a reasoning, coaching configurator from sailsrep.ai, running on SWIFT's real product rules: a language model for the conversation, a constraint solver underneath, and an interactive proposal with the lift, the price and the customer's priorities at the end.
During the Sweden launch, quote requests went from about two a week to about twenty. To be straight about it: a marketing campaign brought the traffic. The configurator is what made that traffic convert. Anders Carlsson at SWIFT put it this way:
"The campaign brought significantly more traffic to our website. The new solution then helped prospective lift customers take the next step, create an initial proposal, and ask to be contacted by a SWIFT salesperson."
The part I find most telling is why SWIFT dared to put it in front of customers at all: every proposal it produces is a lift they can build and price. A language model on its own would happily describe a lift that does not exist. With the solver underneath, it cannot.
How do you measure an online configurator?
Run it as an IT project and it gets measured on availability. Run it as a sales channel and it gets measured on quote requests per hundred visitors, and it gets improved like a landing page: change one question, watch the number, keep what works.
For a reference point, Unbounce's 2024 benchmark puts the median landing page conversion rate at 6.6% across industries. A configurator session and a campaign landing page are not the same thing, so treat that as a yardstick rather than a like-for-like comparison. But if your configurator converts worse than a typical landing page, it is not doing the job of a salesperson. It is doing the job of a brochure.
What does a configurator need to behave like a salesperson?
Five things, in my experience:
- Questions about needs, not part numbers. Ask what the buyer is lifting, where and how often, and derive the specification. That is guided selling, and it is the difference between serving experts and serving everyone else.
- Hard rules underneath. A constraint solver that removes impossible combinations as the buyer goes. Our guide to the CPQ configurator explains why this matters more than the interface.
- A price at the end. Not "we will get back to you". A priced proposal the buyer can forward to their boss tonight.
- Availability at 21:47. On a phone, in their language, without a login.
- An honest exit. When the request is genuinely unusual, the configurator says so and routes it to a person, with everything already captured.
Does a self-service configurator replace the sales rep?
No, and Gartner's own data says so. In May 2026 Gartner reported that 69% of B2B buyers turn to sales reps to validate AI-generated insights. Buyers are not done with reps. They are done with reps as document producers. The companies that get this right let the machine produce the quotes and let people do the judgment: exceptions, negotiation, trust.
Where does your industry stand? Our CPQ maturity map places twenty industries by how far their configurators have moved toward the buyer. And if you want to feel the difference rather than read about it, try today's demo on your phone.
Full disclosure: cpq.se is a Tacton reseller and implementation partner, and we work with sailsrep.ai as a reasoning layer on top of Tacton CPQ. I also work at sailsrep.ai.
Frequently asked questions
Related posts
Simplifying complex configurations, speeding up the sales, reducing errors, empowering customers, and providing valuable market insights. Stay tuned!
Discover how self-service CPQ is revolutionizing personalized customization, simplifying complexity, accelerating sales, enhancing customer satisfaction.
Learn how this transformative tool is reshaping the industry and providing businesses with valuable insights for strategic growth. Stay tuned for more.
