HARMONY IN CHAOS

Our analytical tools transform numbers into knowledge to provide management with efficient solutions, optimizing available resources, for the benefit of people, organizations and the planet.

We develop mathematical models that allow us to know each complex phenomenon by studying the correlations between the variables that influence it, to find harmony in chaos and formulate new growth strategies.

UTILITY

ANALYSIS OF CORRELATIONS

We develop optimization programs for the urban waste collection service, through the study of correlations between the driver’s driving style score and consumption, maintenance costs, type of roads and driver’s profile.

TOKEN

We analyzed the scoring system of the driver, which acquires in real time the accelerations to which the vehicle is subject and which, at each stop, processes these accelerations using an evaluation algorithm. The driver has been assigned at any time with a score that can go from 0 to 10 TOKEN: following the recording of sudden accelerations, the number of tokens is increased; when, instead, a moderate driving style is maintained, the number of tokens decreases progressively.

Red lines denote traits in which there is a worsening of driving style, vice versa, blue lines identify traits in which it improves. The lines corresponding to journeys that did not lead to improvements or worsening of driving style, the vast majority, are highlighted in black.

RESULTS

SAVINGS
OF FUEL

IMPROVEMENT
OF THE DRIVING STYLE

UTILITY

PROCESS OPTIMIZATION

Through Artificial Intelligence technologies we help Utilities to optimize the waste-to-energy district heating system, increasing the efficiency of the waste combustion process.

WEAK CORRELATIONS

To increase steam production in the waste-to-energy boilers, a mix of Artificial Intelligence algorithms was applied with the aim of obtaining the best combination of the various energy sources necessary to maximize the overall efficiency of the plant and to stabilize the process of combustion, ensuring compliance with all the operational constraints of the TLR system. This solution made it possible to automate manual operations and minimize human error.

In this matrix, called the “Data Framework”, the relationships between the variables that influence the WTE (waste-to-energy) combustion process are represented, depicted in different colors and sizes: the red identifies the direct correlations, the blue the inverse, and the yellow the non-linear ones called “weak correlations“, which the human brain struggles to identify. Thanks to the study of these weak correlations, it was possible to provide operators with an effective tool for controlling and optimizing the entire process.

RESULTS

+ 1% INCREASE
OF PRODUCTION
OF STEAM TO PARITY
OF BURN OFF

– 20% REDUCTION OF FLOW RATE
STEAM AROUND
SETPOINT

INCREASE
SUSTAINABILITY LEV
OF THE WHOLE WTE PLANT

BACK OFFICE

ROBOTIC PROCESS AUTOMATION

We develop Robotic Process Automation solutions for repetitive back office activities with low added value, in order to make processes more efficient and free up useful human resources.

MEDUSA

The credit recovery process consists of a series of interventions aimed at obtaining the payment of debts by private parties. The procedure provides for the daily ingestion of data concerning the default status of the entire customer base and the processing of the actions to be undertaken based also on the historical past of each individual customer. This subdivision, previously performed manually, through repetitive “copy-paste”, within 1 week, was speeded up thanks to the application of a developed RPA bot (Robotic Process Automation).

This technology has made it possible to transform operators from executors to process supervisors, thanks to in-depth knowledge of the same, the forecast of future arrears and overall optimization.

RISULTATI

PROCESSING
TIME REDUCTION: 

FROM 7 DAYS TO 25 MIN

INCREASE OF THE ACTIONS PLANNED: FROM 50 BY HAND TO 270

100% OF THE DATA
HISTORY RELATED
TO THE PRACTICE FLOW

+ EFFICIENT KPI FOR
PROCESS MONITORING

INDUSTRY

PRODUCTION PLANNING

By developing Scheduling algorithms that can predict and plan the optimal production sequence, we help companies maximize productivity and reduce planning time.

METRONOME

Through the development of Scheduling algorithms, able to predict and plan the ideal production sequence for the fleet, we help our customers to increase the efficiency of their production processes.

In the time horizon to be optimized, for each instant of time, the Scheduler assigns the production of an article to a press. The aim is to maximize productivity by optimally choosing the order of the items, respecting the orders ordered and minimizing the associated energy consumption.

The Scheduler must look for solutions to this problem in a space with millions of solutions, subject to tens of millions of constraints. The image represents the graph of the setup times of a press. The vertices of the graph are the articles that can be produced on the press and the arcs that connect two vertices are representative of the setup times to pass from one article to another. The thickness of the arc is proportional to the length of the tooling.

RESULTS

REDUCTION
OF PRODUCTION TIMES

GREATER PRODUCTIVITY
WITH THE SAME RESOURCES

REDUCTION OF ENERGY CONSUMPTION

REDUCTION OF RAW MATERIAL WASTE

INDUSTRY

SAP OPTIMIZATION

We help large companies
to increase their internal security level, developing Artificial Intelligence models for ERP control.

NEBULA

By analyzing the data on the SAP system it was possible to know the history of transactions carried out by each individual employee within the company. Through the application of AI algorithms, distinct employee communities were identified according to the type of transactions performed, to each of them the related job role was associated, with specific qualifications on the SAP system. This has favored a downsizing of the total number of user groups recognized by the software and a better management of the permits.

In the graph the dark dots represent the transactions connected to the users who made them, the colored ones represent the users and the different color is indicative of the various communities to which they belong. All the points that share the same color are thus associated with the same job role (ie with the same qualifications on SAP).

RESULTS

OPTIMIZATION
OF SKILL MANAGEMENT

TIME REDUCTION
OF SKILL MANAGEMENT
REORGANIZATION

GREATER SECURITY WITHIN THE COMPANY

UTILITY

PLANT MANAGEMENT

The processes involved in Wastewater Treatment Systems are complex and difficult to predict. The use of a controller based on predictive Artificial Intelligence logics has allowed to optimize the control on the plant.

SPARKLING

The heart of an activated sludge plant is the oxidation compartment where the bacterial populations responsible for the purification treatment are present in the form of flakes held in suspension through the air insufflation. Taking advantage of predictive logics of Machine Learning, the controller understands the dynamic nature of the system and regulates the optimal oxygen value according to the forecast of total nitrogen. This optimization is repeated every 5 minutes, in which the Artificial Intelligence models are interrogated, until the optimal solution is reached.

The graph shows the correlations of the main process variables, every 30 minutes, over the course of a day: from oxygen (process optimization / control variable) to ammonium and nitrates (output from the oxidation tank). The colors of the various correlations highlight their intensity. The histogram above each variable shows the precise value represented by that variable over time.

RESULTS

– 5,1% TOTAL NITROGEN REDUCTION

– 10,4% ENERGY CONSUMPTION REDUCTION

TRADING

ENERGY TRADING PREDICTION

Thanks to advanced mathematical models, we integrate traditional technical analysis information with new correlations between data and new KPIs, in order to identify more effective and profitable strategies for participation in the day before market (MGP), intraday (MI) and the dispatching service market (MSD).

+ 15% INCREASE
OF FORECASTING
MODELS ACCURACY

– 5% REDUCTION
OF FALSE POSITIVES

AUTO

ANOMALY DETECTION

We improve the effectiveness of quality control by applying image analysis algorithms, which automatically identify and report changes in the texture-painting of the cars, or possible structural defects.

+ 85% SUCCESS
IN DETECTING
IMPERFECTIONS
ON MICROMETRIC SURFACES

HELP TO REDUCE
RE-WORKING HOURS

UTILITY

PREDICTIVE MAINTENANCE

We develop predictive maintenance programs for the Utility branch, aimed at improving the forecast of gas losses in distribution networks. We help to optimize the interventions planning on the territory, favoring resource-savings and a lower environmental impact.

+ 34% ACCURACY
IN THE FORECAST
OF LOSSES

ROUTING PLANS OPTIMIZATIONS
OF MAINTENANCE VEHICLES

SPORT

IMAGE ANALISYS

We help sports clubs to define effective and advantageous sponsorship strategies, developing quantitative metrics level enhancement of media visibility of logos on the official shirts, through the application of machine learning algorithms.

MORE
EFFECTIVE
SPONSOR 
STRATEGIES

GDO

INTERNAL COMFORT INCREASEMENT

We develop artificial intelligence models aimed to improve comfort level inside the stores, by applying algorithms that act on the BEMS (Building Energy Management Systems), or through continuous excellent adjustments of the HVAC systems (“heating, ventilation, air conditioning and cooling”).

– 9.8% REDUCTION
OF ENERGY CONSUMPTION
IN 1 YEAR WITH THE SAME ASSET, IN 10 STORES

INCREASE OF INTERNAL COMFORT LEVEL

Our analytical tools transform numbers into knowledge to provide efficient management solutions, optimizing available resources, to support people and the planet.

UTILITY

ANALYSIS OF CORRELATIONS

RESULTS

FUEL SAVING

TIMELY IMPROVEMENT OF DRIVING STYLE

GREATER INVOLVEMENT AND MOTIVATION OF DRIVERS

UTILITY

PROCESS OPTIMIZATION

RESULTS

+1% INCREASE OF STEAM PRODUCTION AS A BURN REFUSAL

– 20% REDUCTION OF STEAM FLOW RATE AROUND SETPOINT

INCREASED THE SUSTAINABILITY LEVEL OF THE WHOLE WTE SYSTEM

BACK OFFICE

ROBOTIC PROCESS AUTOMATION

RESULTS

REDUCTION OF PROCESSING TIME: FROM 7 DAYS TO 25 MIN

INCREASE OF PLANNED SHARES: FROM 50 MANUALS TO 270

100% OF THE DATA HISTORY RELATED TO THE PRACTICE FLOW

KPI + EFFECTIVE FOR PROCESS MONITORING

INDUSTRY

SUPERVISED SCHEDULER

RESULTS

REDUCTION OF PRODUCTION TIMES

INCREASED PRODUCTIVITY AS WELL AS RESOURCES

REDUCTION OF ENERGY CONSUMPTION

INDUSTRY

SAP OPTIMIZATION

RISULTATI

OPTIMIZATION OF ABILITY MANAGEMENT

REDUCTION OF THE TIME OF REORGANIZATION OF THE SAME

GREATER SECURITY WITHIN THE COMPANY

UTILITY

PLANT MANAGEMENT

RISULTATI

– 8.1% TOTAL NITROGEN REDUCTION

– 16% ENERGY CONSUMPTION REDUCTION

TRADING

EFFECTIVE FORECASTS
FOR ENERGY TRADING

+ 15% GREATER ACCURACY
OF FORECAST MODELS

– 5% REDUCTION
OF FALSE POSITIVE

AUTO

ANOMALY DETECTION

85% SUCCESS IN DETECTING IMPERFECTIONS ON MICROMETRIC SURFACES

 HELP TO REDUCE RE-WORKING HOURS
DUE TO PRODUCTION ERRORS

UTILITY

PREDICTIVE MAINTENANCE

34% ACCURACY IN THE FORECASTS OF LOSSES

OPTIMIZATION
OF ROUTING PLANS
OF MAINTENANCE VEHICLES

SPORT

IMAGE ANALISYS

+ MORE EFFECTIVE
SPONSORSHIP STRATEGIES

GDO

INTERNAL COMFORT INCREASEMENT

– 9,8% CONSUMPTION REDUCTION

INCREASE OF INTERNAL COMFORT LEVEL