Home › Case studies › Valfonta
Success story · Industry · customer service
Valfonta: they stopped answering the same question a hundred times
Their technical team spent the day resolving the same questions. Today an assistant answers them in seconds, in the small hours too, and only what really needs it reaches a person.
At Valfonta, clients' technical questions were eating the team's time, and those arriving outside office hours went unanswered until the next day. We put an artificial intelligence assistant on their website that has learned from the company's own documentation: it answers instantly, at any hour, and alerts a person when the enquiry deserves it.
24/7
attending to clients, at night too
−55%
repeated questions to the team
<1 min
to give an answer
+30%
enquiries captured outside office hours
* Metrics and testimonial pending confirmation with the client.
The context
Valfonta sells technical products, and that has a consequence: before buying, the client asks. Dimensions, materials, compatibilities, lead times. Legitimate questions, almost always the same ones, that ended up on the desks of the people with the most knowledge in the company.
Whoever buys a technical product does not buy on impulse. They have a project, a machine or a refit under way, and they need to be sure the part fits before signing anything. That is why they ask: they are not haggling, they are looking for peace of mind. And until they have it, they do not move.
What is more, they almost never ask in one place only. They open three or four websites of similar suppliers and put the same question to all of them. From there the decision is played out on very simple ground: whoever answers first and best starts with an advantage, because they convey that they know their subject.
The result was twofold: a technical team trapped in email and a pile of enquiries coming in during the evening or at the weekend that were not answered for many hours. In a sale where whoever answers first wins, those are opportunities going cold.
The challenge
When you look closely, the problem was not a lack of information: it was that all that information depended on somebody going to look for it and writing it out again. Each answer cost a few minutes, but the minutes were multiplied by dozens of emails a day.
- The same questions, over and over, taking hours from those who add the most value.
- Outside office hours, nobody answered: and the client does not wait.
- All the information existed, but it was spread across documents and product sheets that had to be gone and looked for.
- Enquiries that were lost simply by arriving late.
The most expensive cost was the one that appeared in no accounts: interruption. An engineer preparing a complex quotation who stops three times to answer a question about a dimension does not lose three minutes, they lose the thread. And picking it up again costs considerably more than the email takes.
The wear and tear counted too: repeating the same diameter for the tenth time motivates nobody. And meanwhile, Saturday morning's email was still there, waiting for Monday, with a client who by then had already asked somebody else for a quotation.
The solution: an assistant that already knows its subject
We did not put in a generic assistant of the kind that answers anything. We put in one that only talks about what it knows: Valfonta's products and documentation. The difference looks small and it changes everything, because whether the client trusts the answer depends on it.
1 · Putting in order what they already had
We gathered and put in order the technical documentation, the product sheets and the most common questions. That is the material the assistant learns from: the company's own, verified information, not invented.
It was the most laborious part and also the one that left the most value on its own: reviewing everything turned up incomplete product sheets and figures that existed only in two people's heads.
2 · Putting it where the client is
The assistant lives on the website, in plain sight. The client asks in their own words, as they would ask a sales rep, and gets the answer straight away.
Nobody has to fill in a form or look for a phone number. Somebody arriving at eleven at night on their phone writes their question exactly as they think it and carries on with the purchase without leaving the page.
3 · Knowing when to call in a person
It solves the repetitive part on its own. When it detects an important enquiry — a quotation, a special case, a large client — it passes it to the team with all the context, so nobody has to start from scratch.
The sales rep does not get an empty alert: they get what the client was looking for, what they were told and where things were left. They can call knowing what the conversation is about, which is the exact opposite of opening with “how can I help you?”.
4 · Making sure no contact is lost
Every conversation is recorded alongside the client's record. What used to be an anonymous question at eleven at night is now an identified contact who can be followed up.
That allows something as simple as calling the next day and saying: “I saw you were asking about this part number”. The client notices they have been listened to, and the company stops throwing away interest it had already paid to win.
5 · Teaching it to say “I do not know”
An assistant that answers everything is a risk. We gave it clear limits: if the question falls outside what is documented, it does not fill the gap with guesses. It says so and offers to put the client through to a person.
It may look like giving something up and it is exactly the opposite. Acknowledging a limit generates more trust than a brilliant but wrong answer, especially when whether a part fits an installation depends on that answer.
6 · Improving it with what clients ask
Each month there is a review of what has been asked, what was resolved well and where the assistant had to pass the enquiry on. With that the documentation is completed and next time it answers on its own.
That review is worth something in itself: it is the real list of the market's questions, written by the clients, and it serves to improve the product sheets and even the sales team's pitch.
What changed day to day
Before, a normal morning began with an inbox full of emails piled up since the previous afternoon. A good part of them were short questions: a dimension, a material, a lead time. Answering them took the first hour and a half of the day, the clearest, the best for thinking.
Now that hour and a half no longer exists as a task. The repeated questions have been answered from the moment they were asked, many of them in the small hours. What the team finds on arrival is a handful of alerts about enquiries that do deserve a person, each one with the earlier conversation in front of them.
The change shows above all in the tone: work is no longer about putting out fires but about choosing who to call first. And when a difficult question comes in, it can be given the time it demands.
Results
The client gets an answer instantly, at any hour. The technical team has got a good part of its time back and devotes it to what really requires a person. And the company no longer loses the enquiries that arrive when the office is closed.
The repeated questions reaching the team have been cut to less than half. They have not disappeared, and that was not the goal: the ones still coming in are the ones that add something, the ones a professional has to handle because there is a nuance, a volume or a decision involved.
Office hours stopped being a boundary. A significant share of enquiries happens outside the working day, when the client finally has a quiet moment to look at suppliers. That stretch used to be blank; today it is selling time like any other.
And there is an effect that is not easy to measure: peace of mind. Knowing that nothing is left unanswered and that every contact is recorded lifts a small, constant worry.
Why this case matters
- The knowledge you already have in documents can work for you twenty-four hours a day instead of sleeping in a folder waiting for somebody to open it.
- Answering fast sells: in technical selling, the first to answer starts with an advantage, even if they are not the cheapest on the list.
- Properly contained artificial intelligence takes work off the team without answering more than it should: it knows its subject, and when it does not know, it alerts a person.
What it includes
It is not a closed product you plug in and that is that: they are pieces assembled to fit each company, starting from the documentation that already exists and ending with measurement. This is what was set up at Valfonta.
An AI assistant on the websiteIt learns from their documentationAnswers 24/7An alert to the team when it is dueA record of every contactMeasuring results
“Before, there were questions that went unanswered until the next day. Now a client asks at eleven at night and has the answer straight away. And when the enquiry is one of those that needs a person, it reaches us with all the context: we know what they need before we pick up the phone.”
— Valfonta's management
Frequently asked questions
What if the assistant invents an answer?
It is the first question everybody asks, and rightly so. That is why we do not use a generic assistant: it answers only from the company's documentation and its scope is contained. If something falls outside that, it does not improvise: it alerts a person.
In technical selling, a wrong figure is not a minor error: it can end in a part that does not fit and an angry client. That is why we would rather it answered fewer things and that all of them were true.
Do clients not prefer to talk to somebody?
They prefer to be answered. For a dimension or a lead time, an immediate answer beats waiting until the next day. And when the enquiry is important, the assistant hands over to a person, it does not block the way.
In fact something curious happens: many people ask things they would never write in an email, because they seem too basic. Those small doubts are often the ones that hold a purchase back.
Is it a lot of work to maintain?
Little. It is updated when the products or the documentation change, just as you would update a catalogue. The bulk of the work is the setting up: putting in order the information you already have.
From then on a periodic review of what clients ask is enough: a while each month, and it is usually done happily, because ideas for the website come out of it.
How long does it take to go live?
It depends above all on the state of the documentation. If the product sheets and catalogues are reasonably well ordered, it moves quickly. If scattered information has to be rebuilt, that phase rules the rest of the schedule.
What we do aim for is that it starts working early with a reduced scope, rather than waiting months to have everything perfect. Better that it answers the ten most frequent questions well and grows from there.
Is it any use if my products are very specific?
The more specific the product, the better it works. An assistant that talks about everything adds nothing; one that knows a particular technical catalogue in detail solves exactly what nobody else can.
The condition is having the information written down somewhere: product sheets, manuals, compatibility tables. If it is only in two people's heads, the first step is getting it out of there, and that is worth doing anyway.
Can it serve clients in several languages?
Yes. It understands the question in the language it is written in and answers in that same language, with the same documentation behind it. For a company that exports, that means serving a foreign client without depending on who happens to be in that day.
It is worth reviewing how the technical terms come out in each language, because that is where credibility is decided. It is a short review and it is done once.