Big Data & AI Paris 2025 — Feedback: Exclusive Networks Automates Sales Forecasting with Alteryx
On October 1 and 2, 2025, Prime Analytics took part in Big Data & AI Paris, one of Europe’s leading events dedicated to data and artificial intelligence.
Alongside our strategic partner Alteryx, our teams led several demos and workshops showcasing real-life use cases in Data Quality, AI, and automation.
Among the highlights of the event: our case study workshop with Exclusive Networks, dedicated to automating sales forecasts with Alteryx.
Feedback workshop with Exclusive Networks
Presented by Lancelot Seners (Head of Business Controlling — Exclusive Networks) and Radhi Toujani (Data Scientist — Prime Analytics), the workshop entitled “How to Forecast Sales: the Alteryx ‘Sales Predictive Forecast’ Workflow in Action at Exclusive Networks” drew a large audience on Demo Stage 4.
On the agenda: field feedback, time savings for sales teams, a test-and-learn approach, seasonality management, and an ARIMA/ETS model comparison — all handled autonomously, without depending on the central IT team.
The goal: to achieve an accuracy of around ±3% and deliver actionable forecasts for management and regional teams alike
Video — Watch the full workshop replay
Background and objectives of the project
Lancelot Seners, Head of Business Controlling at Exclusive Networks, opened the session by presenting the group and the context of the project.
A global player in cybersecurity, Exclusive Networks operates in 45 countries with more than 3,000 employees and an annual turnover of around 6 billion euros. The company plays a key role in connecting security solution providers with major technology integrators.
Faced with market complexity and fragmented commercial data, the group sought to modernize its sales forecasting process.
“The dream of any sales team is to push a button and instantly know how much they will sell in the coming months.”
The predictive forecasting project was born from this ambition — to make short- and medium-term financial planning (3, 6, 9, and 12 months) more reliable, while reducing the time local teams spend updating their forecasts.
The goal was clear: achieve less than ±3% deviation between automatically generated forecasts and actual sales, to justify large-scale adoption within the company.
Technological choices and approach
Exclusive Networks was already using Alteryx for its analytics needs. The solution was a natural choice for this project thanks to its no-code approach and its ability to combine data manipulation, modeling, and automation.
The group partnered with Prime Analytics to strengthen the project’s technical dimension and accelerate the implementation of statistical models.
“The value of Alteryx is simplicity. But as soon as you move into predictive modeling, you need real expertise. That’s where Prime Analytics came in.”
The first step was to create a simple Proof of Concept (PoC): an Alteryx workflow defining data sources, transformations, and automatic forecast generation.
Once this foundation was validated, the teams ran hundreds of tests comparing ARIMA and ETS statistical models across different horizons (1, 3, 6, or 12 months) and levels of aggregation (group, region, deal size, etc.).
Optimizations and learnings
A second phase of the project introduced additional parameters, including:
- • consideration of exchange rates,
- • and activity seasonality, with a marked sales peak in December.
These adjustments significantly improved model accuracy, especially for quarterly and annual forecasts.
“The more stable and larger the dataset, the more reliable the model. Where activity is regular, we are now below a 3% deviation.”
Lancelot also highlighted a key point: the importance of business knowledge when processing historical data.
Before training the model, some periods had to be adjusted to exclude exceptional events (loss of a major client, acquisitions, etc.) to ensure the reliability of the projected trends
Implementation and automation
To scale up the solution, Exclusive Networks invested in an Alteryx Server.
This allows workflows to run automatically at scheduled times (overnight or on weekends), greatly reducing the workload and execution time for teams.
“If the model had run on a single workstation, it would have taken weeks. With the server, we were able to run more than 300 forecasts almost continuously.”
Thanks to this automation, forecasts are now generated without dependence on the central IT team — a major advantage for the agility of controlling teams.
The contribution of Prime Analytics
-
- The contribution of Prime Analytics was key in:
- • configuring and calibrating the ARIMA and ETS models,
- • establishing a collaborative process between data scientists and business teams,
- • and training internal teams to help them become autonomous
- The contribution of Prime Analytics was key in:
- “What would have taken us several months to do alone, we achieved in just a few days with Prime Analytics. Their Alteryx expertise and support saved us a lot of time.”
Technical demonstration
-
- Radhi Toujani, Data Scientist at Prime Analytics, detailed the technical structure of the Alteryx “Sales Predictive Forecast” workflow.
-
- The process consists of six main steps:
- 1. Data cleaning and preparation,
- 2. Transformation to the format required for the models,
- 3. Checking the quality of the input data,
- 4. Training of ARIMA and ETS models,
- 5. Generation and comparison of forecasts,
- 6. Visualization of results and selection of the best model.
-
-
- Comparative analysis showed that the ETS model provided a better representation of seasonality and monthly variations than the ARIMA model for Exclusive Networks’ data.
Results and prospects
- The project made it possible to:
- • automate the production of forecasts on a global scale,
- • reach a average accuracy less than 3%,
- • execute more than 300 automated forecasts,
- • and strengthen the autonomy of the finance and controlling teams.
- The project made it possible to:
- This pragmatic approach demonstrates Alteryx’s ability to make complex predictive models accessible to non-technical teams — while delivering measurable gains in reliability and productivity.
We will get back to you as soon as possible.
Please try again or contact us at contact@primeanalytics.fr.