AI Assistant
An assistant one click away on any screen, answering questions about how the system works by consulting the official documentation, in the user's language.
The questions it handles are the usual ones: how to issue a credit note, where to change a price list, what a particular status means, or why an option is not showing. Questions with a written answer somewhere, currently resolved by interrupting a colleague or opening a support ticket.
The effect shows in two places. In the learning curve for new joiners, who ask instead of searching a two-hundred-page manual. And in support volume, which drops because most tickets were usage questions rather than incidents.
It answers with the context of the screen the user is on, so there is no need to explain where you are for it to understand what you are asking.
How-to questions
The questions that today interrupt a colleague
In every company there are three or four people who have become the software's living manual. They get asked in the corridor, by phone and by message, and each interruption costs twice: the person asking, who was stuck, and the person answering, who was doing something else.
The assistant answers those questions from the official documentation:
- How to issue a credit note.
- Where to change a price list or a special price.
- What a particular document status means.
- Why an option is not showing in the menu.
It is one click away from any screen and answers in the user's language, with the context of where they are: there is no need to explain which part of the software you are in for it to understand the question.
Where it shows
The learning curve and support volume
New joiners
Somebody starting takes weeks to move comfortably around a management system, and during those weeks they depend on somebody else having time. With the assistant they ask when the question arises, at the moment they have it and without feeling they are being a nuisance, which is what stops many questions being asked at all and leaves them to trial and error.
The support team
A significant share of the tickets any software provider receives are not incidents: they are usage questions with an answer already written in the documentation. Resolving them the moment they come up cuts the volume and, above all, the time a user spends blocked waiting for a reply.
Smart FAQs complete this with semantic search over the FAQ base, finding the answer even when the question is phrased differently or contains typos.
How far it goes
About the software and, if you enable it, about your data
The base assistant answers about how the system works. That is public product information: it does not access your company's data and does not need it to explain how something is done.
When you also want to ask about your own figures (how much have I invoiced this month, which work orders are still open) it extends with AI over your data, which is a separate module with its own permission control: each user only gets the information their profile allows.
The separation is deliberate. A help assistant available to the whole workforce is one thing; access to business information, which requires deciding who can query what, is another.
Clearing things up
Frequently asked questions
Answers to what we get asked most.
What kind of questions does it answer?
Questions about how the system works: how to issue a credit note, where to change a price list, what a status means or why an option is not showing. It consults the official documentation and answers in the user's language.
Do I have to tell it which screen I am on?
No, it answers with the context of the screen it is opened from. It is one click away anywhere in the software, which is what makes people ask at the moment the question arises rather than later.
Does it access my company data?
The base assistant does not: it answers about how the product works, which is public information, and does not need your data to explain how something is done. Querying your own data is a separate module, "AI over your data", with its own permission control.
Does it reduce support tickets?
Yes, because a good share of tickets are not incidents but usage questions with an answer already written in the documentation. Above all it cuts the time a user spends blocked waiting for a reply, which is the cost nobody invoices but everybody pays.
Artificial intelligence and OCR
AI reads your invoices and creates the document, and an assistant answers in plain language about the software and your data, respecting each user's permissions.
OCR · Document Digitisation
Upload the PDF or a photo of a purchase invoice, delivery note or order, and the system extracts the supplier, line items, net amounts, taxes and due dates, and creates the document for you.
View moduleAI over your data
Extend the assistant to query your company's real data (sales, customers, stock, work reports, clock-ins, outstanding payments) by asking in plain language.
View moduleSmart FAQs
Semantic search over the FAQ knowledge base: the assistant understands what the user means, not just the exact words they typed.