Intelligent mobile assistance and product recommendation service

Technology leadership - how to produce a real value product recommendation system for your customers with Artificial Intelligence and Data Mining.

SDART has been approached by representatives of the NTT Group with a request to consult and suggest future directions for their technology development. NTT Group is managing all area of telecommunication and ICT business. This includes the mobile operator NTT DOCOMO. It also has some financially consolidated arms in EMEA such as Dimension Data and Cirquent.

Problem faced

The EMEA financial representative office of Japan's NTT (Nippon Telegraph & Telephone) Group contacted us with a request for a concept solution for an intelligent mobile assistance service. The aim for the project was to discover new possibilities in particular regarding the improvement of their CRM (customer relations management) strategy used by the current “I-concier” system. The main focus was concentrated on the possibility of developing a new intelligent service that would offer real value suggestions and assistance to the customer.


SDART’s technical expertise were required for these project included:

  • Data Mining - Customer Profiling & Categorization
  • Artificial Intelligence Systems - automatically creating recommendations for customers


We selected the most relevant Data Mining processes.

Sales and user interaction:

  • Customer interest profiling
  • Purchase prediction (know what and when a customer is most likely to purchase)
  • Sales improvements / conversions rates
  • Sales forecasting
  • Predictive stock control
  • Customer / product profiling and matching
  • Estimating the best cost value for the product/service based on a customer profile
  • Churn analysis
  • Cross Selling and Up Selling
  • Propensity Modelling e.g. to respond to a mailing / purchase of a product
  • Contact strategy development
  • Feedback analysis

Social media response analysis:

  • Use of social media as a source of data for analysis
  • Crowd sourcing and predicting interest based on social media


  • Customer profiling
  • Marketing campaign profiling
  • Marketing ROI increase
  • Conversion rate increasing
  • Segmentation and Targeting Analytics
  • Customer Lifetime Value

E-commerce analysis:

  • Web statistics analysis, what attracts costumers and how to retain them
  • Product grouping/recommendation according to the customer profile, interests and shopping habits


Solution proposed:

SDART created a detailed concept vision and explanation of the possible applications of the mobile advisory system (I-concier example). Also included was an explanation of the advanced Data Mining strategy that should be applied for successful customer profiling and recommendation processes. The planed enhancements of the systems functionality had to stay in line with the business-defined objectives. The technical advantage was to be achieved by delivering a sophisticated recommendation algorithm based on a flexible functionality of the system and various data sources and predictive modeling of the customer.

automated product recommendation service

Our proposal consisted of three main parts:

  • Data gathering - specific to the customer requirements
  • Discovering and analysing of the data while transforming it to pertinent information
  • Information delivering

The key technologies in the advanced Data Mining processing include:

  • Automated Data Analysis, intelligent filtration and formatting (Decision Tree, AI methods, Neuro-networks, K-Means Clustering)
  • Customer behavior Modeling techniques (Gaussian Mixture Models)

Functions performed by the systems intelligence:

  • Classification
  • Estimation
  • Prediction
  • Affinity Grouping or Association Rules
  • Clustering
  • Profiling