IBM Cloud Pak for Data

IBM Cloud Pak for Data quickly and easily combines all data sources and data storage systems relevant for analysis into a single user view. This eliminates time-consuming and costly transformation and migration processes.

The Watson Knowledge Catalog supports the subsequent organization, classification, and governance of data with policies and rules. With the Watson Knowledge catalog, companies develop an information architecture that precisely addresses their analytical and information requirements.

The catalog enables a “shop-for-data,” supports users with a glossary, and thus represents an essential basis for the extended use of data. In addition to these governance aspects, which can extend up to a data origin analysis (data lineage), simple data transformation/ETL also takes place on this level.

Analytics services analyze and visualize extensive data sets based on the information architecture created. For example, companies create Cognos dashboards or use the SPSS modeler for data mining and statistical analysis. The AutoAI service is part of the service portfolio: Artificial intelligence develops automated and self-developed AI models that are best suited for analyzing the data sets available in each case. This is a task that usually takes weeks.  AutoAI systematically tests a wide range of AI models systematically, makes many adjustments automatically, and in the end makes recommendations about which AI model will succeed the fastest for the respective task.

In particular, the model created can also be exported as Python code and used further. This saves even experienced data scientist a lot of time and less experienced data scientists have a good basis for making a start. Overall, AutoAI allows even companies with less or no data science skills to use and deploy artificial intelligence in their own company.

With Watson Machine Learning, artificial intelligence becomes an integrated component for challenging data analyses and can thus be integrated directly into business processes. This allows for the operationalization/industrialization of AI. In addition to operating AI models, other aspects play an increasingly important role:

• Data drift: IBM Cloud Pak for Data automatically detects when a model should be reworked because the data on which it operates changes fundamentally (e.g. data before and during the Corona crisis).
• Fairness: It is ensured that no injustices affect model decisions, e.g. men are favored, or young people experience disadvantages.
• Ability to explain the decision: The decisions of a neural network can be analyzed for the main drivers, which increases acceptance of the AI model.
• Transparency of the decision: It is possible to understand which parameters have been incorporated and which AI model has been used for each and every AI decision made. This, too, increases acceptance and represents a solution approach in the current discussion about AI ethics.

These four aspects, which contribute significantly to the successful operationalization of AI models, are fully addressed in the Cloud Pak via the Watson OpenScale service.

A rapidly growing company has acquired several business units and wants to integrate the new partners into a uniform IT architecture as quickly and efficiently as possible. With IBM’s data virtualization layer, it is possible to efficiently manage many local data pots, systems, and reports in different locations without having to migrate the data sources.

However, the construction of a classic enterprise data warehouse is also possible without any problems. The IBM Cloud Pak for Data speeds up the necessary ETL processes (Extract, Transform, Load) and reduces costs. The analysis of the data and visualization of the results are subsequently performed by specialist users with IBM Cognos analytics as an optional component of the Cloud Pak for Data.

A company wants to use self-learning artificial intelligence to improve its forecasts and put its decisions on a more solid, data-based basis. However, the company does not employ any experts such as data scientists who are very familiar with the new technologies.

With the AI self-service AutoAI, even specialist users without special AI knowledge can determine whether and if which AI models makes sense and are successful in their specific use case. Specialist users select the relevant datasets, AutoAI independently tests the relevant AI models, and then automatically provides a strengths-weaknesses analysis at the end of the process. Creation takes only a few minutes. The result can be put directly into production and integrated into existing or new business processes.

A wholesaler wants to unify the work of its data analysts and integrate it into its business processes. With IBM Cloud Pak for Data, various roles such as data scientists, business analysts, data engineers, controllers, business analysts, and app developers work transparently on the same platform. The end-to-end integration of processes and roles enables faster and greater penetration of AI across the enterprise.

Scarce resources, which have to be optimized according to availability, costs, or benefits, are available in every company across all sectors, be they personnel, materials, or capital. IBM Cloud Pak for Data optimizes the use of operating resources, quickly reducing costs, double allocation, and penalties.

Here’s an example: The marketing department of a large service provider would like to invite you to a workshop. Selected customers are supposed to receive a fine bottle of wine in advance. With the help of Cloud Pak for Data, marketing analyzes, among other things, the size of the company, the position, the personal preferences of the guest, and the customer’s history, to determine which customers are worth the investment and what return on investment can be expected on which level. In addition, there are what-if simulations for better planning in sales, marketing, production, and finance.

IBM Cloud Pak for Data uses the conversational AI built into the Watson Assistant to conduct dialogues in natural language. Chatbots generate several benefits: They are ready to go 24×7, they increase customer satisfaction, and relieve qualified personnel who can devote themselves to more important tasks. The Watson Assistant is available in multiple languages and can be integrated into internal corporate systems such as CRM or ERP. The integration into the Cloud Pak for Data thus allows the operation of the chatbot in the customer’s environment and not as a service from the public cloud. Many customers have discovered that this is an advantage: Access to the data becomes easier and more secure. In addition, further customer-specific dialogues can be designed.