One profile per user
Language, field, tone and preferences live in one place. Every part of your product asks the engine instead of guessing.
Your product, speaking each user's language.
The communication engine sits between your product and its users. It learns who each person is — their words, their field, how much they know — and adapts everything your product says to them: chat, e-mails, error messages, forms. And it turns what they say back into tasks your product can do.
We are building it now. Write to us if you want to use it.
Contact usYour users are different. Why does your product sound the same?
A beginner and an expert get the same screens, the same e-mails, the same error messages. One finds them too technical, the other too basic — and both have to learn your product's words instead of using their own.
How does it work?
You start by giving it what it needs to know: your fields, the kinds of users you have, the words you use. It is built from your own documentation, not written by hand. From then on, it works in both directions:
-
Your user says something
In their own words: "make it more playful", "how is my site doing?" The engine understands them.
-
It becomes a task
The engine turns it into a task your product can do, and your product does it.
-
Your product has news
Something happens — a trial is ending, a payment failed — and your product decides it is worth saying.
-
The engine says it their way
It writes the message in the user's language, tone and level, in chat or by e-mail.
-
It learns
Every conversation teaches it more about each user, so the next message fits better.
01 says
02 does
03 reports
04 explains
05 learns
Why one voice?
Your product talks to people in many places. When each place has its own idea of who the user is, your product sounds like several different people. The engine is the one place that knows each user, so everything sounds like one voice that knows them.
Answers from your own truth
When a user asks something, the engine answers from your documentation. People outside your company only get what is public.
No AI to build yourself
You don't need your own conversational AI, user models or tone rules. That is what the engine is.
What your company knows can be its one source of truth
Everything the engine needs comes from your documentation. We keep ours in the knowledge engine, and each of us works with it through this engine, in our own words — maybe it fits the way you work too.
If you build software too, it can take requests in plain words
We use it in our own development engine: each of us describes what we need in our own words, and it turns that into a clear spec to work from — see if it would work for you too.
Who owns what it learns?
Your users' data stays yours
You stay in charge of what it learns about your users, kept apart from every other product that uses the engine.
Nothing learned leaks out
What it learns in a private conversation is never used when it talks to people outside.
Not trained on your users
Conversations with your users don't train the engine for anyone else.
Still have questions?
Not yet. We are building it. If you want to use it, write to dive@brightly.in.
No. Chat is one channel. The engine adapts everything your product says — e-mails, error messages, forms — and remembers each user.
A stable identity for each user, documentation the engine can read, the tasks it may ask your product to do, and the events worth telling users about.
Not really. It needs to know who it is talking to; without that it would be a plain FAQ bot.
At the moment we use it in our own products. We're putting it on top of our knowledge and development engines, so each of us can work with them in our own words. In our website builder, it's the guide that answers questions and suggests what to do next, in plain words.