A Roadmap to Intelligent Business (part II)

by Mike Ferguson

March 2005

Stages to Implementing Intelligent Business

There are four main stages to implementing and intelligent business strategy. These are:

1. Define an intelligent business architecture
2. Integrated business intelligence (the BI element)
3. Enterprise Business Integration (the operational element)
4. Intelligent Business (the combining of operational and BI)

Note that stages 2 and 3 can and should be done in parallel.

The intelligent business architecture (Figure 1) acts as a blueprint for implementing the intelligent business. This architecture brings operational and analytic applications together in the context of business processes that present content to users via a personalised, secure web-based user interface – an enterprise portal.

Integrated business intelligence (stage 2) involves standardising on a common BI development platform to reduce the complexity and provide common technology for BI development. This stage also involves repairing existing BI systems by introducing a shared business vocabulary (data names and definitions) across all BI applications to support common understanding and consistency. Having done this, BI can be integrated so that islands of custom and package based BI systems plug into CPM scorecards and dashboards for rock solid enterprise corporate governance.

For operational performance management, companies need to capture operational events by integrating ETL data integration software with application integration platforms (EAI) to collect events and trigger on-demand near real-time data integration of structured and unstructured content when specific events occur. In addition, event driven on-demand automated analysis and a common rules engine will provide support for business activity monitoring (BAM), on-demand recommendations and alerts for guiding operations personnel towards achieving business objectives in the intelligent business.

Enterprise business integration (stage 3) should ideally be done alongside stage 2 so that teams in both areas work together and plan where and how to integrate BI into operations. Based on this architecture and achieving specific strategic objectives, this is about implementing integration at four main levels:

1. User interface integration,
2. Business process integration
3. Application integration
4. Data integration

so that these integration initiatives are done in unison to achieve a common strategic business objective e.g. to reduce operational cost.

User interface integration is done using enterprise portals with built-in collaboration and content management. Enterprise portal technology allows a company to give customers, suppliers, partners and employees personalised access to integrated content

Business process integration involves the separation of business processes from applications so that integrated business processes can be separately modelling (using business process modelling tools) to guide the execution of business process across multiple applications inside and outside the enterprise. A process engine then manages process execution. Once underway, executing business processes can then be monitored using business activity monitoring (BAM) and activity based costing (ABC).

Application integration sits below business processes and involves the use of a common platform for application integration and the use of web services, UDDI, SOAP, WSDL, and XML. The process engine that is executing a business process sends industry standard SOAP XML messages to application web services connected to the enterprise integration platform to carry out specific process activities.

Stage 4 is the intelligent business stage. This is the integration of business process operations and BI for full business performance management. Here BI web services provide integration of BI into operational applications to make processes intelligent. In this way, each process activity leverages the relevant BI and/or recommendations on-demand. In addition the rules of a business process that describe the path that a process takes can be driven by BI. Hence if customer intelligence indicates a ‘gold’ customer, then they might be led one way though a process whilst a ‘bronze’ customer might be led another way through the process. This is made possible by common rules and a rules engine shared across applications. Vendors like PegaSystems, CA, Microsoft (Biztalk 2004), and Fair Isaac offer such engines and they are already doing this with their customers’ business processes. In the world of intelligent business, the rules engine uses rules to cause automated alerts and recommendations during business activity monitoring (BAM) and alerts executives via CPM scorecards. In addition, because this rule driven decision/action engine is itself a web service it can be integrated into operational applications to issue recommendations on-demand to guide operational users using operational applications. Also in intelligent business, BI is integrated with enterprise portals for intelligent personalised e-business and guided operations (figure 4).

Note that a business process can span multiple organisational departments and multiple applications across the enterprise. So it is not enough to just understand the process. It is necessary to understand the roles of people who participate in the process and the applications they use in each activity so that we understand what BI is needed and what we have to do to integrate it into business operations (figure 5).

Clearly some activities in a process are performed automatically by software while others are performed manually by people. In automated activities BI can be integrated via web services so that a program can request BI on demand via a standard mechanism. If the activity is performed by a person, then several other things need to be understood. These include the role of the user and what applications people use when performing a specific business process activity. Also if the person is normally mobile they would need to have access to BI from a mobile device? It is highly likely that people in many different roles throughout the enterprise can all contribute in some way to achieving the same business objective. Role recognition is therefore extremely important and different approaches may be needed to integrate BI into business processes to fit with the role of each user (e.g. customer facing call centre operator, bank branch counter staff, salesperson etc.). The objective is to deliver the right BI in the context of a specific process activity being performed at a specific time. People in multiple roles contribute to the same objective. Therefore, detailed investigation needs to determine the following:

• What process tasks they perform and what applications they use
• During what tasks is BI needed?
• What BI do they need to help them contribute to the common objective?
• In what form do they need BI e.g. reports, guided analytics, instant live recommendations integrated into another application, alerts…..
• Do they have time to use a BI tool or not?
• Do they need the BI delivered on a mobile device?
• Does the use of BI systems need to be totally transparent to the user? i.e. automated analysis, automated recommendations, automated alerts etc. 
• What actions does a person in this role need to take?
• Do they need to collaborate with others before taking action?
• Is the action expected to be automatic (i.e. no people)?

Answering to these kinds of questions will lead to a clear understanding of what kind of closed-loop BI integration strategy is needed to support specific users who are performing specific activities as part of a business process.  It should also highlight that each role may need a different closed loop BI integration strategy. For example a customer service representative in a call centre has no time to use a BI tool and must have BI integrated into the operational application they use to guide them during dialogue with customers if they are to ever become more effective in contributing towards a strategic business objective. Equally an executive needs BI integrated into CPM software. Both require BI to be integrated in different ways to help them do their job in an intelligent business. Identifying the correct BI integration strategy the fits the user need (e.g. call centre operators have no time to use BI tools) is therefore a critical success factor.

Technical Requirements For BI Integration

Over and above the investigative work defined above, the following are a non-exhaustive list of technical requirements that help integrate BI into operational systems.

• Integrated business intelligence
• Integration between EAI and ETL for event driven near-real time data capture
• XML input support in the data integration platform (ETL) technology
• Web services support from BI tools and analytic applications
• Common data naming for the same data across all BI data models and BI tools
• CPM software to build dashboards and scorecards that link lower level metrics to KPIs and objectives in the business strategy
• An automated rules engine integrated with BI to help manage and drive day-to-day business operations

Based on the above needs there are a number of ways to integrate BI into the enterprise:

1. Integration of analytical applications with operational applications using an enterprise portal for access and exploitation by internal and external users
2. Embed analytics in operational applications during application development
3. Introduce web services to dynamically integrate analytical processing with internal and partner operational applications
4. Deploy real-time processing for user alerts, on-demand recommendations, and automated actions

Conclusions

There is no doubt that intelligent business is coming and companies will strive to turn their organizations into intelligent businesses. A shared business vocabulary is critical to achieving this and to creating consistency across the enterprise. Also business integration is needed using both a standard enterprise integration platform and a standard business intelligence platform. In addition business intelligence systems need to be “cleaned” up to make them consistent before integrating BI with operational business processes.