We endeavor to package the enabling technology in the form of problem-specific, configurable, extensible platforms leading to several other benefits such as (i) quick derivation of custom solutions through easy configuration and/or extension (ii) operationalization of past experience and expertise of the organization (iii) enabling juniors to function at the level of experts.
To realize this broader vision, MDO is currently investigating three streams, namely Organizational Decision Making, Model Driven Regulatory Compliance, and Mining Enterprise Models. Each of the three streams is briefly discussed below.
Organizational Decision Making
The research seeks to reduce the reliance on human-based processes in order to support decision-making that leads to organizational change. We believe that the scale and complexity of modern organizations, combined with the required pace of change, make human-based processes incomplete and prone to error. In addition, such processes introduce a knowledge management problem where normal staff turnover leads to the loss of key knowledge.
The research premise is that information about the who, what, how, and why aspects of an enterprise is necessary and sufficient for organizational decision-making activities under various operating conditions. The program proposes that a single, unified language or specification to formulate these aspects in a localized relatable form can help overcome present problems due to multiple partial views that need to be integrated for deriving insights. Such a simulatable specification can enable what-if and if-what scenario playing, thus leading to data-driven and informed decision making.
Model Driven Regulatory Compliance
Various financial crises in the last decade have forced regulatory authorities to take increasingly sterner actions against organizations that fail to show compliance. Organizations thus need to devise effective and efficient means to enact compliance.
Setting regulatory expectations always involves some form of interpretation, consultation, negotiation, and risk trade-off. Operationalizing these changes is complicated by the fact that terms used in the text of regulation and in the operation details are different, but target concepts that mean the same. It should be possible to carry out the change activity with regard to operational practices with ease, when complying with new regulations or with amendments in existing regulations.
The research hypothesis is that systematic and structured mapping of vocabularies of regulation texts and operational specifics, along with purposive models of regulations and business objectives, form a complete compliance solution. The primary lines of investigation include:
- Reducing semantic disparity between regulations and domain-specific enterprise operational specifics
- Modeling the vocabularies of regulation texts and operational specifics and mapping them incrementally
- Modeling changes in regulations
- Ensuring that such models enable preparedness of the enterprise to change their operational specifics in compliance to regulations
- Modeling business and regulation objectives in a holistic manner
- Ensuring that such models enable protection of business objectives despite regulatory changes
- Evaluating models for insights by which both regulatory compliance and business objectives can be achieved together
Mining Enterprise Models
Research on organizational decision-making and automated regulatory compliance expects information about enterprises in a structured form, that is, a set of models. For example, in an enterprise, one would like to know who are key stakeholders and why; what are the key tasks they perform; and how do they realize these tasks to meet the desired objective. The who, why, what, and how aspects of an enterprise can be alternatively viewed as a system of actors (who) exchanging messages constituting of goals (why), events (what), state and behavior (state transitions or how), thereby establishing an implicit relationship between them. Once the sources of structured and unstructured data are mined in the light of the above aspects, they can be stored in a model repository for subsequent question-answering and decision-making.
Although the research stream has a broad focus, the current emphasis is to mine information from unstructured knowledge sources only, in particular, from unstructured, document intensive business processes. This information will help automate the business processes, identify bottlenecks, and even redesign the processes if necessary.
This group is headed by Kulkarni, Vinay
Individual investigations streams under MDO are led by Souvik Barat (Organization Decision Making), Sagar Sunkle (Regulatory Compliance) and Suman Roychoudhury (Mining Enterprise Models).