On your website
It answers what people ring you about today: opening hours, terms, delivery times, how to book an appointment. If someone wants to book or talk to a person, it takes them there.
AI chatbots that cite their source
An assistant that answers from your data, tells you where each answer comes from and, when it doesn’t know something, admits it and puts you through to a person. For your website, your WhatsApp or your team.
The one on this website works like that. Ask it.
Where it works
On WhatsApp Business
Your customers already message you on WhatsApp. The assistant handles the usual questions and takes orders with the details needed, so your team only has to confirm them.
Inside your team
An internal assistant that looks things up in your documents and your management software. “What price did we sell that RCD to this customer at?” “What does the maintenance contract say about call-outs?” It answers and shows you the document or the figure it took the answer from.
Same question, two answers
A contractor asks an electrical supplies distributor: “Have you got the 40 amp RCD? How much will it cost me?”
A generic AI replies: “40 A RCDs protect against earth leakage. The price depends on the manufacturer and the distributor; I recommend checking with your usual supplier.”
Correct, and no use at all. It knows nothing about your warehouse.
An AI with your data replies: “Yes. The 2P 40 A 30 mA type A RCD is €40.02 on the trade price list, before your discount. There are 47 units in stock. Source: price lists and stock as of 30 September.”
That’s the example from our distributor demo: fictional company and data. The difference isn’t the AI model. It’s that the second one looks up your data and says where it got it.
When it doesn’t know, it says so
An assistant that makes up what it doesn’t know is worse than no assistant. Ours is built to answer only with what it has and, if it doesn’t have it, to say so:
“I don’t know that and I’d rather not make it up. Shall I put you in touch with a person?”
From there it offers to book a call, write on WhatsApp or leave a message for your team. The conversation doesn’t end in a dead end.
Transparency and the law
Anyone talking to your assistant has to know they’re talking to an AI. Article 50 of the EU AI Act requires it, and we comply from the first message.
- A notice at the first interaction: it’s an AI, not a person, and it can make mistakes.
- What is kept and for how long, stated clearly. On this website we keep conversations for 30 days, without the IP address, and you can ask for yours not to be kept.
- A person always within reach.
We explain it in detail in how we use AI and in the privacy policy, both in Spanish.
Examples by industry
- A distributor: the sales rep checks price and stock from the van. Retail and distribution
- An accounting firm: the internal assistant finds the client’s document and the rule that applies. Accounting and law firms
- A clinic: it answers questions about opening hours and helps patients prepare for a visit, without ever touching a diagnosis. Clinics and healthcare
- A rural guesthouse: it replies in Spanish and English at two in the morning. Clubs, hospitality and tourism
Questions about assistants
Does it make up answers?
It’s built not to: it answers from your documents and your data, cites the source and, if it can’t find the answer, says so. Even so, an AI can get things wrong. That’s why we test it with you, using real questions, before putting it in front of your customers.
What happens to what my customers write?
We decide with you what is kept, where and for how long, and we explain it to whoever is writing. If nothing needs keeping, nothing is kept.
Does it answer in other languages?
Yes. It replies in the language people write in. We test it with you in the languages you need before it goes live.
Can it do things as well as answer?
It can take an order, suggest an appointment or leave a note for your team. Whatever it does, it does under the same rules as an automation: anything doubtful is checked by a person.
Shall we talk about your assistant?
Tell us what your customers or your team ask you every day. That’s enough for us to know where to start.