how Axolotl works, our predictive AI engine
how Axolotl works, our predictive AI engine
how Axolotl works, our predictive AI engine
Axolotl is ExcellerIA's proprietary artificial intelligence engine.
It turns your data into actionable signals to anticipate changes, allocate your resources at the right time, and adjust your strategic decisions — without heavy redesign.
Axolotl is the proprietary artificial intelligence engine of ExcellerIA.
It transforms your data into actionable signals to anticipate variations, allocate your resources at the right time, and adjust your strategic decisions — without heavy overhaul
Axolotl is the proprietary artificial intelligence engine of ExcellerIA.
It transforms your data into actionable signals to anticipate variations, allocate your resources at the right time, and adjust your strategic decisions — without heavy overhaul
Nos clients
Nos clients
Nos clients
Why is the Axolotl different?
An AI designed for business
Axolotl builds on your internal tools (ERP, CRM, CMMS) and your other data flows (websites, data lakes, ...) and automatically enriches itself with external data (weather, seasonality, market trends, competition, ...).
It detects weak signals invisible to the naked eye, anticipates behaviors, and evolves with your business context, ensuring a concrete ROI from the first weeks.


An AI designed for business
Axolotl builds on your internal tools (ERP, CRM, CMMS) and your other data flows (websites, data lakes, ...) and automatically enriches itself with external data (weather, seasonality, market trends, competition, ...).
It detects weak signals invisible to the naked eye, anticipates behaviors, and evolves with your business context, ensuring a concrete ROI from the first weeks.

A quick and simplified integration
No major IT projects. Our solution adds as a module to your Business Intelligence tools (even the oldest ones), and can be operational in a few days.
The installation is smooth, and the usage remains accessible, even without a dedicated data team.


A quick and simplified integration
No major IT projects. Our solution adds as a module to your Business Intelligence tools (even the oldest ones), and can be operational in a few days.
The installation is smooth, and the usage remains accessible, even without a dedicated data team.

A continuous learning and integrated compliance
Axolotl is designed to meet the most stringent regulatory requirements:
• Transparency and explainability of algorithmic decisions
• Data protection built in from the design phase
• Responsible machine learning that refines its models with each new piece of data


A continuous learning and integrated compliance
Axolotl is designed to meet the most stringent regulatory requirements:
• Transparency and explainability of algorithmic decisions
• Data protection built in from the design phase
• Responsible machine learning that refines its models with each new piece of data

The data dilemma
The data dilemma

Dirigeant
j’aurai besoin d’une analyse complète pour lundi : des projections à six mois sur les volumes, les marges et les risques.
Est-ce faisable dans ce délai ?

Dirigeant
j’aurai besoin d’une analyse complète pour lundi : des projections à six mois sur les volumes, les marges et les risques.
Est-ce faisable dans ce délai ?

Analyste
Je croise les ventes réelles, les prévisions budgétaires, les tendances du marché. Je prépare un fichier clair d’ici vendredi.

Analyste
Je croise les ventes réelles, les prévisions budgétaires, les tendances du marché. Je prépare un fichier clair d’ici vendredi.
Analysts manage the data. Executives know the business. How do you align both perspectives fast enough to generate a reliable forecast, without weeks of back-and-forth?
With Axolotl, data and business expertise operate together. The engine goes beyond computation: it absorbs your real-world constraints (commercial context, supplier shifts, exceptional operations) to stay aligned with your market dynamics.
You no longer have to hope your figures match what actually happens. Axolotl’s continuous learning delivers forecasts that mirror the reality to come.
Analysts manage the data. Executives know the business. How do you align both perspectives fast enough to generate a reliable forecast, without weeks of back-and-forth?
With Axolotl, data and business expertise operate together. The engine goes beyond computation: it absorbs your real-world constraints (commercial context, supplier shifts, exceptional operations) to stay aligned with your market dynamics.
You no longer have to hope your figures match what actually happens. Axolotl’s continuous learning delivers forecasts that mirror the reality to come.
Analysts manage the data. Executives know the business. How do you align both perspectives fast enough to generate a reliable forecast, without weeks of back-and-forth?
With Axolotl, data and business expertise operate together. The engine goes beyond computation: it absorbs your real-world constraints (commercial context, supplier shifts, exceptional operations) to stay aligned with your market dynamics.
You no longer have to hope your figures match what actually happens. Axolotl’s continuous learning delivers forecasts that mirror the reality to come.

How Axolotl Works in 6 Steps
Preparation and enrichment of data
Processing by our engine
01
Data collection
Gathering and secure sending of your internal data (ERP, CRM, etc.).
01
Data collection
Gathering and secure sending of your internal data (ERP, CRM, etc.).
01
Data collection
Gathering and secure sending of your internal data (ERP, CRM, etc.).
02
Quality control
02
Quality control
02
Quality control
Cleaning, formatting, preparation to ensure the reliability of the results.
03
External enrichment
03
External enrichment
03
External enrichment
Adding data (weather, seasonality, behaviors, location...) for maximum accuracy.
04
Training
04
Training
04
Training
learning phase (deep learning) with explainable models
05
Automated hyperparameter tuning
05
Automated hyperparameter tuning
05
Automated hyperparameter tuning
Fine-tuning drives accuracy, but it’s traditionally slow and manual. Axolotl automates the process, evaluating hyperparameters at scale and retaining the most efficient. This boosts speed and raises forecast quality.
06
Provision of predictions
06
Provision of predictions
06
Provision of predictions
Preparation and enrichment of data
Processing by our engine

The technological forces of Axolotl
Proprietary AI
Axolotl relies on an internally developed engine, controlled from end to end.
We use open source frameworks (TensorFlow, PyTorch), without depending on third-party services.
This architecture guarantees total sovereignty over the processes, suitable for sensitive data and regulatory requirements.

Proprietary AI
Security and compliance
External factors
Processing capacity
Smooth installation
Pre-trained models
Frugal AI
Proprietary AI
Axolotl relies on an internally developed engine, controlled from end to end.
We use open source frameworks (TensorFlow, PyTorch), without depending on third-party services.
This architecture guarantees total sovereignty over the processes, suitable for sensitive data and regulatory requirements.

Proprietary AI
Security and compliance
External factors
Processing capacity
Smooth installation
Pre-trained models
Frugal AI
Proprietary AI
Axolotl relies on an internally developed engine, controlled from end to end.
We use open source frameworks (TensorFlow, PyTorch), without depending on third-party services.
This architecture guarantees total sovereignty over the processes, suitable for sensitive data and regulatory requirements.
Proprietary AI
Security and compliance
External factors
Processing capacity
Smooth installation
Pre-trained models
Frugal AI
Concrete results, fast.
Example: At a major retailer, Axolotl helped identify in advance a sales activity that exceeded the processing capacity of the planned staff.
Result: loss of revenue avoided, adjustment of the schedule, customer satisfaction, time and margin saved.
Axolotl was designed for those who want to act faster, decide more accurately, without heavy investment.













