LTR Intelligence

An assistant that doesn’t just answer.
It helps you get things done.

LTR Intelligence understands your spaces, your documents and your tasks. It answers, it prepares, and it only acts with your consent.

ConsideredBeing designed: no LTR Intelligence feature is available today. This page presents the direction of the product.

Answer first, then prepare the action.

In this preview, the assistant answers a question from your information, then offers to draft the follow-up message. Nothing is sent without your approval.

Three levels of help

Answer, prepare, carry out.

The difference between an assistant that answers and one that acts comes down to one question: who presses the button? In LTR One, it’s always you for anything that commits something.

  1. Assistant

    It answers.

    You ask a question, it searches your spaces and answers with its sources.

    “What did I promise Sam last week?”

    Considered
  2. Copilot

    It prepares.

    In a document, a task or an email, it suggests a version, a summary or a list of steps. You keep, edit or discard it.

    Summarise a meeting and turn it into tasks.

    Considered
  3. Supervised agent

    It carries out, with your consent.

    It chains several steps for you, but stops before any action that commits: sending, deleting, inviting, paying.

    Prepare this week’s follow-ups, then wait for your approval.

    Considered

What it will be able to do

Truly understand your information.

  • Contextual memory

    Considered

    It remembers your projects, your preferences and your recent exchanges, and you can review and erase what it keeps.

  • Search and synthesis

    Considered

    A question in everyday language, an answer that pulls together information scattered across your notes, documents and tasks.

  • Understanding your data

    Considered

    It recognises a deadline in an email, a task in meeting notes, a document linked to a project.

  • Local and cloud AI

    Considered

    The chosen principle: simple processing locally when possible, the rest with a provider under contract, with no training on your data.

The rules it follows

  • It acts with your permissions, checked by the server at every action, never more.
  • It cites the source and date of the information it uses.
  • It does nothing irreversible without your approval.
  • It can be switched off: LTR One’s core features don’t depend on AI.

Frequently asked questions

How is it different from a chatbot?

A chatbot answers from general knowledge. LTR Intelligence will work from your own information, with your permissions, and will be able to prepare actions that you approve.

Is my data used to train a model?

No. It’s a design rule: the providers chosen will have to commit not to train their models on your data.

When will it be available?

We won’t give a date until it’s ready. Beta members will be the first to try it.

LTR One is coming. Be among the first.

We’re opening the beta gradually, by invitation. Request access, or leave your address to hear about the launch.