AUTHORS

Anna Tarantini
Data Visualization and
Engineering Expert
@Bip xTech

In planning and control, one of the typical issues is the huge amount of effort to get reliable monthly closure data. The typical calendar of official closures terminates only 1 or 2 weeks after end of month, due to long processing in ERP or other financial reporting systems. So, P&C often builds own monthly closing tools (typically MS Excel based), kind of  “digital twins” of business logics with PxQ modeling, implying huge effort each month.

This effort also prevents these mechanisms being used before end of month to determine pre-closing estimates.

This article explains how Bip xTech, through a mix of AI, RPA and Data engineering, addressed the  questions coming from top management, asking for faster, prompt and timely actual P&L data closing, and reliable and frequent predictions on end of month closure within a short term (during month): we called it “Smart Closing”.

The “Smart Closing” solution is a forecasting tool with a busines logics engine enabling calculation of the P&L for multiple service lines (e.g. in Telco it was used for consumer, business/SME, enterprise, both fixed-line and mobile).

Its engine starts from the granular actual quantities and pricelists data, forecasting quantities for the rest of month with AI, mixing them up with PxQ and management control logics, and following attribution to P&L lines and bottom up aggregation to construct the entire P&L.

The final output is shown on a BI dashboard with different views (executive and operational) and KPI scorecards.

All the process, from the RPA data gathering, to the forecasting, to the calculation and aggregation, up to the final BI data preparation, lasts only a few hours, allowing the exercise to be repeated even daily, when necessary.

Advanced analytics to support monthly pre-closing

 Traditional Planning and Control models reason on reporting cycles on a monthly basis, where actual information gathering is often tied to «official closure» which prefer «completeness» rather than «prompt answer». Typical forecast models, where present, are normally managed at an aggregated level (overall P&L line time series forecasting) and only partially can intercept low signals from the market.

Bip xTech solution is based on two founding principles: accuracy and simplicity.

Accuracy is guaranteed by the processing of daily historical data of main low level indicators which contribute to determine the value of commercial results (Service Margin), by the analysis of the relevant detailed dimensions (e.g. prices/tariff plans, offers /promotions, customer segments, usage type, …) which can influence trends and projections for each commercial scenario and by the availability of advanced analytics models able to interpret KPI historical curves and evaluate the evolution of each commercial scenario

Simplicity is ensured by the automation of information gathering and consolidation, the capability to make predictions for single specific scenarios (e.g. based on offer type, client profile, etc.) and the immediate analysis of results and trends comparison.

The whole process begins with the RPA ENGINE which automatically extracts data from the different sources in a secure way (IT auditable) but without the need of IT implementation.

Then, in the Data Preprocessing layer, heterogeneous set of data is read, cleaned and prepared based on technical and business rules at every run. Data Quality checks are performed, any data issue is automatically notified.

For each granular KPI, Machine Learning Forecasting Models are trained automatically (Data Modelling

Layer) every month on most recent data leveraging on Auto-ML techniques generating the forecast. This means that underlying AI stratus self retrains every month choosing the best model for each of the thousands of time series forecasted, guaranteeing best performance and low errors.

Subsequentlygranular forecast (P’s and Q’s) is aggregated using business rules to generate P/L lines. Model performances are evaluated to auto-adjust the algorithms on the next run (Output Producer)

The Smart Closing portal acts as a web application orchestrator because various functions have been implemented to allow users control, autonomy and above all simplicity in the run activities of the predictive models. Users are appropriately profiled to allow appropriate role management and access to the administration functions.

Through the web portal, the user can independently and at any time:

  • Upload files (price lists, catalogs, dates, etc.) to support the forecast
  • Run the RPA engine to retrieve data from Legacy systems
  • Run the preprocessing engine
  • Consult the control dashboards, output of the previous point (data freshness, output check, column check, BPC check)
  • Run predictive models (AUTO-ML)
  • Insert manual adjustments to the forecasts
  • Launch the BI dashboard update
  • See the «preview» and «official» versions of the BI dashboards

With the delivery of this use case, Bip allows the final user to completely autonomously get an accurate prediction of the monthly closing, with a short term time window, based on the actual granular data collected and uploaded to the Smart Closing portal.

The solution is available on cloud, leveraging the best AI Cloud native functionalities for better performance and precision.

Success Story

Smart Closing tool is currently being adopted by a major Telco company to support the monthly fast closing / pre-closing process.

By using the technology, the company registered a progressive reduction of prediction error down to less than 1% for runs performed at day 28th of the month, 2-3% error for runs early in the second part of the month.

Manal effort for data extraction and calculation has been reduced to zero (a few hours to control input and output), thanks to the use of robotic automation and AI.

Bip xTech practice

Bip “xTech” Practice supports clients in the adoption path of edge Technologies among which Data/AI/Cloud/Blockchain/5G/Hyperautomation to optimise, boost efficiency, automatise decision processes and evolve business through new tech enabled services.

We act across the whole adoption funnel, since Strategy, to Design and Implement, up to Operations, accompanying our customers also in the transformation process.

Bip xTech has international coverage: we operate all over the world starting from our primary offices in Italy, UK, Spain and Brazil.


If you are interested in learning more about our offer or would like to have a conversation with one of our experts, please send an email to [email protected] with “Smart Closing” as subject, and you will be contacted promptly.

Read more insights

red lines

Get in touch

Milan, Italy | BIP xTech Head Office

Torre Liberty Building
Galleria de Cristoforis 1, Milan, 20121

[email protected]

    This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.