SententIA sits inside the back office, in the “Ask SententIA” panel. Whoever is working there writes what they need in plain words, and the assistant works out for itself which of the three jobs it is.
Every read of data goes through the back office on the session of the person who asked: the assistant has no route of its own. Its abilities are granted per profile, in the permissions grid you already use.
Ask it where a setting lives, why a case is not doing what you expect, or to take you there and set the work up.
Where a setting lives: the answer is the on-screen path, in the names you read. Why a case is not doing what you expect: which condition is not met. Or it takes you there and sets the work up.
If the person asking cannot open a screen, neither can the assistant; if their scope is one site, the assistant reads that site. It has no route of its own to the data.
Three families: knowledge searches the manuals, platform reads back office data, screen reads the open page, fills it in and presses. They are granted in the same grid as every other permission: nothing new to learn.
On request it opens the right screen, fills in the fields and presses the buttons for you, and you see all of it and confirm. A profile without the screen family gets an assistant that answers but never touches.
Two menu entries translate the catalogue’s markets and selections in bulk, into the languages you choose, and a selection travels with its market: “1” under Match Result and “1” under First Goalscorer come out as two different words.
The engine is chosen in configuration, one for the whole back office. One of the engines can run on dedicated hardware, for operators who must keep the data in house. Changing engine is a setting, and nothing else moves.
It sees what the person talking to it sees, and nothing more: every read goes through the back office on that person’s session. If their scope is one site, it reads that site. The conversation dies with that session.
No. The assistant works on the session of the person talking to it: what that person cannot open, or cannot touch, it cannot do either. When it cannot, it says so, and says what it can do.
Yes. The assistant’s abilities are permissions like any other, granted per profile in the same grid: knowledge, platform, screen. A profile without the screen family gets an assistant that answers but never touches.
Two menu entries translate the catalogue’s markets and selections into the languages you choose, in one go. The trade glossary and examples already approved guide the wording, and you read the result through before you save it.
We wrote it ourselves: SententIA is not a separate tool, it is the “Ask SententIA” panel inside the back office. The engine that answers, though, is chosen in configuration, and changing it moves nothing else.
Send us a couple of lines: the markets you want to open, the products you need, and when you want to go live. You’ll get a straight answer on what we can do, and what we can’t.
