Business intelligence (BI) and data warehousing solutions have primarily been about gathering data from various applications and combining it to provide useful information for decision-makers. This approach presents only half the picture because it is data-centric, lacking the context of business process in which the data was created. Mapping of underlying process information with BI data enables business users to understand why specific results have been achieved, as well as which processes produce desired results and which do not. Such information is critical to effective decision-making.
This article shares our experience in executing a proof of concept (POC) by integrating business process management (BPM) with BI. The approaches and methodologies proposed are based on real-time integration experiences.
Building BI into business processes is a crucial step that enables organizations to create smarter procedures by leveraging the analytical data available in the data warehouse. The benefits of doing so include greater agility, effectiveness and efficiency. For example, if an organization wants to automate the campaign process to target its existing customers, it can leverage the intelligence available by designing a smarter process which accesses and utilizes BI data in real time to make decisions on the fly.
In the same way, the insurance underwriting process provides relevant data to BI as an event occurs. This captured data is further analyzed and used for creating real-time dashboard solutions providing critical key performance indicators (KPIs).
Therefore, it becomes important to integrate BPM processes seamlessly with BI and provide relevant data of importance when creating such an environment.
A quick comparison of a traditional BI implementation versus the way BI ought to be implemented to support a full-fledged BPM initiative (BI for BPM) has been outlined in Figure 1 (see PDF below).
In a typical BPM environment, BI provides support by:
- Setting up high-level metrics with drilldown analytics,
- Sending out alerts when performance deviates from planned goals,
- Turning alerts into automated procedures for escalation and remedial action in real time and
- Sending distilled data back to the BPM process for performance improvement.
In some critical business scenarios, data from BI assumes far more importance than process data in analyzing and optimizing business processes. For example, if we take a real-time customer credit score to calculate loan interest rates, BI harnesses transactional data from a business process and matures it with understanding so it can lead to action. In this instance, the order fulfillment process can leverage BPM/BI integration in the following ways:
- Should we process this order even though it exceeds the customers credit limit?
- Should we escalate this order to regional sales manager because the credit score is low and this customer is a priority client?
- Should we extend more credit to this customer even though he or she has outstanding payments?
Thus, integrating process data and BI data enables decision-making in context and in real time.
While working on the integration of BI and BPM, remember that BI information must be combined with BPM for processes to be used for timely decision-making. Ideally, one should ensure that a unified architecture with BI and process engines sharing the same schema is available so that organizations can take action in real time. Data captured through the BPM process can be utilized for strategic and tactical reporting or operational BI reporting.
Service-oriented architecture (SOA) is the infrastructure which IT uses to enable business initiatives. BI creates a business face for data, and BPM creates a business face for the process, while SOA is the underlying approach taken for building applications across the business.
Figure 2 (see PDF below) depicts the typical steps involved while implementing a BI integrated BPM solution.
The architecture of an end-to-end BPM-integrated BI solution is shown in Figure 3 (see PDF below). The proposed architecture publishes BI services that are consumed by the BPM process. As the data gets generated by BPM process, a BI service is called to capture the relevant process attributes and then stores them in the BI database. Dashboard monitoring and reporting operational KPIs reflect data with minimum latency. A Windows service senses the arrival of new data and triggers extract, transform and load (ETL) to load data in the dimensional model for further analysis and aggregation.
Ways to Integrate
There are three approaches in which BPM and BI solutions can interact with each other and deliver value: BI on process, BI embedded in the process and event-driven process.
BI on Process
In this approach, BI uses BPM process data for dashboard and analytics. There could be two possible scenarios: synchronous and asynchronous.
In the synchronous process, the relevant data is passed on to BI synchronously when the appropriate transaction is completed. In order to implement synchronous integration, Web services are recommended. A dashboard solution for monitoring operational performance is a representative example of this approach.
In the asynchronous process, data is pooled to BI as an end-of-day process, and SQL connects are recommended. A dashboard solution for tactical and strategic management pooling and aggregating data from multiple sources is an example of this approach.
Detriments and benefits:
- BI utilizes BPM data on an as-needed basis.
- Synchronous methods of implementing this approach reduce latency.
- Services needed from BPM processes must be realized by BI and exposed for getting consumed.
BI Embedded in the Process
BPM processes are normally rich in information concerning process and transaction within its scope but poor in terms of external information. If the external data required is static, BPM can get the same results using its built-in integration functions. In case the external data required is complex and the data source is the output of the analysis done by BI, then BI process must be invoked from BPM. For example, if a bank wants to give the customer the best rate for a loan, the bank would need info about the customers credit history, profitability and risk profile. A BI process can be configured for this and can be invoked from BPM within its decision rules to get the required data. Web services are the preferred method for such implementations. Here the BPM process uses the information and data collected in BI for decision-making purposes.
Detriments and benefits:
- As a part of overall process, BI analytics is embedded in the workflow. Thus, there is optimal utilization of BI capabilities.
- This approach is best suited to cases where the process decision requires immediate inputs from BI. In overall terms, it increases process efficiency and improves productivity.
- Malfunctioning or error in BI results impacts BPM process adversely.
Processes help organizations and people deal with tasks that need to be performed, providing a consistent and repeatable way of handling them. Event-driven processes are those where the dashboard outputs of BI process trigger a process in BPM. For instance, the dashboard on critical stocks going below the inventory level (scorecard goes red) can result in a BPM process being triggered, which would send emails to the appropriate authorities, initiating a particular workflow. Organizations using BI event management techniques can help individuals focus on critical aspects of the process and build confidence that nothing is missed. BPM workflows can be invoked via Web services from BI.
Detriments and benefits:
- Action is initiated immediately as the event occurs.
- Latency is reduced in the overall system. Stakeholders are alerted immediately, especially when results deviate from set goals.
- It requires more effort to manage different events and their subsequent planned actions.
Current Industry Trends
BPM primarily aims at automating decision-intensive processes, while BI products lack the capacity to define and execute processes around decisions. This gap is now starting to close as vendors from both sides start to create more numerous and deeper partnerships and integrations.
The alliance between BPM and BI provides far more benefits than detriments. In any organization, the process is the DNA, and efficiency and uniqueness are often key competitive differentiators. Stringent regulatory compliance today forces organizations to look into operations and build processes that allow process experts to analyze for inefficiencies as well as to build decision-making capabilities directly into processes or allow for closer event-based exception handling. These are the key areas where BI and analytics play a critical role.
An enterprise can realize the best ROI when BI and BPM are brought together and integrated on the same platform, which enables joint development, management and integration. The following points are critical to decision-making while deciding on such investment:
- Up to what degree are BPM and BI components integrated?
- Does the platform support tools in everyday use, like Microsoft Office?
- Does the solution leverage existing investments or IT assets?
BPM/BI integration framework establishes a bridge between all data-generating units and data-gathering units in the company, regardless of location or types of data stored. Thus, an integrated world emerges where interprocess communication and sharing is the core, finally driving corporate success.
At the ground level, there are complexities in implementing this framework, but industry trends show that strong alliances and collaboration between BPM vendors and BI vendors are taking place. Such a move will bring maturity and build efficiency in building BPM-integrated BI solutions.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access