Discover Curebot Assistant.

Knowledge. Solutions. Inspirations.
Optimize the value of your resources with AI.

Our vision of Curebot Assistant.

Interpret the diversity and creativity in every result!

Our aim is to make intelligence, analysis and decision-making more effective and accessible to all. We're moving up a gear in this direction. Curebot Assistant is built on a self-hosted Open Source LLM model that lets you exploit the knowledge hidden in information.

We are committed to offering transparency and flexibility to our users in their use and adoption of Curebot Assistant. You'll know exactly where to start, so you can discover new avenues of thought, carry out your analysis step by step, and determine which actions to take.

Trust-generating AI.

Our LLM model is self-hosted on our secure, private servers.

We have designed Curebot with the highest standards of security and confidentiality. The development of our AI technologies confirms this commitment. Our professional, self-hosted architecture ensures that our customers' data is always protected, while harnessing the power of AI.

Do you have any questions?
Curebot Assistant answers your questions.

Query Curebot Assistant.

Enter the instructions of your choice to get the answers you need. Just ask.

Analyze resources.

Experience the potential of AI to support analysis and exercise a critical eye on information.

Summarizing and synthesizing.

Discover the key points to remember in just a few seconds to test new hypotheses.

Extract data.

Interpret information to help you make the best decisions.

Streaming responses.

Get instant insights for a more dynamic, personalized intelligence experience.

Prompt library.

Choose from a library of prompts specially designed by watch experts.

Contact us to test Curebot Assistant.

Research and development.

Several months of R&D were spent testing different use cases, calculating quantitative statistics (response time, stability, performance...) and dissecting qualitative data (linguistic, semantic, lexical, syntactic relevance, formatting, omission, hallucination...) with a blind human annotation campaign by multiple annotators.

To frame our analysis and study the various constraints, we designed an experimental protocol. The aim was to compare and differentiate between various LLMs, both proprietary and OpenSource, as well as various infrastructures, hosting providers, backend technologies, etc. As part of this protocol, we tested the combination of all these criteria to obtain results that were as exhaustive as possible. In this way, we've gathered valuable data on the relative performance of each combination, helping us to make informed choices when developing Curebot Assistant.

The diversity and richness of the experiments led us to objectively choose the best solution to implement LLM on a self-hosted infrastructure, and thus meet our customers' needs for the implementation of these generative artificial intelligence technologies within our intelligence platform.

What does the future hold?

Curebot Assistant is just getting started. We dream and work every day to develop new features! We are continuing our research and exploring how AI can be harnessed to help our customers improve their intelligence and analytics experience.

The rest to come soon...