Making Intelligent Decisions in Information Intensive Business Activities
Data driven decision making has been a business goal for many years, growing out of the big data and analytics revolution. With broader interest in artificial intelligence (AI) over the past year or so, new ways to approach decision processes are gaining momentum. For routine business decisions, generative AI and integrated data models are rapidly changing the approach away from basic business intelligence (BI) and analytics solutions and the subsequent dashboards and queries derived from them, to a natural conversation with the data to uncover the necessary answers to support the decision. This move to a natural question and answer process over pre-built dashboards and queries, accelerates the decision process and enables users to find specific answers much more simply and independently instead of relying on either predefined and thus limited views of the data; or IT, data scientists and data analysts for custom views. This is a big step forward for day-to-day decision making, and can add value in more complex business activities, but is, however, not a broad enough solution to solve the more complex decision scenarios.
When you examine business decisions, you could evaluate them based on the information needed (or complexity of the decision) and the importance or criticality of the decision. The business decision model could look something like this:
In at least half of the decision scenarios in the model, the support needed to make a data driven decision is mostly solved by a combination of BI/analytics solutions and the newer generative AI (natural query) based approaches. The exceptions though, are in the highly complex decision categories. Current solutions are relatively inadequate to support employees in these activities. Decisions in these categories require a multidimensional information model to support the involved employees.
Highly Complex and Highly Critical Business Decisions
There are many business decision scenarios that fall into this highly complex category. Here are a few examples:
Major Mergers or Acquisitions:
Complexity: Deep due diligence into the target company's financials, risks, legal standing, market synergies, and cultural fit. Integration planning is extraordinarily intricate.
Criticality: Potential to transform the company or lead to massive losses if strategy and integration fail.
Entering New Markets (Especially International):
Complexity: Analyzing unfamiliar market dynamics, regulatory frameworks, consumer preferences, competitor analysis, and potential supply chain issues.
Criticality: High initial investment of resources. Failure can severely damage reputation and financial standing.
Significant Investments in New Technology:
Complexity: Evaluating return on investment, lifespan of the technology, disruption to current operations, employee training requirements, and compatibility with existing systems.
Criticality: Can make or break a company's ability to compete or innovate. Poor technology choices lead to wasted resources and falling behind rivals.
Large-Scale Product Launches:
Complexity: Multi-faceted execution across R&D, marketing, supply chain, sales, and customer support teams. Thorough market analysis needed.
Criticality: Success significantly boosts revenue; failure damages brand image and incurs massive losses.
Responding to Major Disruptions:
Complexity: Situations like global crises, economic downturns, or supply chain breakdowns require rapid re-evaluation of strategies. Decision-making must balance long-term survival with short-term needs.
Criticality: Existential threat potential. Incorrect responses can doom a business, while successful pivots can open up new opportunities.
Restructuring or Downsizing Operations:
Complexity: Balancing financial necessity with employee impact, severance costs, maintaining essential operations, and legal compliance in the process.
Criticality: Poorly handled situations damage company culture and can trigger long-term reputational harm that hurts recruiting talent.
Important Considerations:
Context is Key: The size of the business matters. A decision that's routine for a multinational corporation might be existentially critical for a small business.
"Information Needs" is Subjective: What is considered "high" information needs depends on leadership experience and access to resources for analysis. A decision might feel complex due to the lack of those things, rather than the intrinsic nature of the problem itself.
Many of these decision types (like M&A, new technology investments and product launches) could be categorized as large-scale project execution. Additional project-specific complex & critical decisions include:
Resource Allocation: Balancing personnel, budgets, and materials across a massive, multi-layered project is an ongoing challenge. Inefficiencies here lead to delays and overruns.
Change Management: Large-scale projects are rife with change requests. Deciding which to accommodate, their impact on timelines and cost, and how to communicate them effectively is crucial.
Risk Assessment and Mitigation: Proactive identification of potential risks (technical, environmental, personnel-related) is complex. But having strategies to minimize their impact on the project is what sets successful project managers apart.
What is an Information Model
An information model is a structured representation of the essential concepts (entities) within a particular area of interest (a domain) and the relationships between them. It provides a blueprint for understanding how data and information are connected, organized, and should interact within a business system.
Think of an information model in these ways:
Language of a Business: Determines how the business talks about its people, processes, locations, policies, and assets.
Foundation for System Design: Acts as a blueprint for creating databases, software applications, decision support systems, and other systems that rely on the business's data.
Complete representation of the business: Customers, employees, transactions, assets; and can be enriched from external sources depending on the specific needs of the business
Governance Tool: Provides standards and rules to maintain clean, consistent, and trustworthy data.
Types of Information Models:
Conceptual Information Model: A high-level, abstract view of the main concepts and their relationships that matter in a business domain. It's focused on the big picture and less on specifics.
Logical Information Model: More detailed than a conceptual model. It includes specific information about data elements (attributes), the nature of their relationships, and sometimes even data types.
Physical Information Model: Takes things to the implementation level by describing how the logical model will be physically stored within a database management system (DBMS).
Here’s a simplistic view of company data invarious “states”:
In a very basic way, the model needs to provide the tools and processes to transform the data in the “problematic state” into the desired state, with the support state as an interim state in the model.
Benefits of an Information Model for Businesses
Clear Understanding of Entities and Relationships: Provides everyone with a consistent understanding of their core business information.
Improved Data Quality: Promotes data consistency thanks to standards and rules established during the modeling process.
Effective System Design: Serves as a solid base for database design and creation of reliable software that maps to real business processes.
Better Decision-Making: Supports AI / ML, automation, analytics and business intelligence because information is structured in a clear, accessible manner.
Using the Information Model
Having a complete information model is the first step in providing an enhanced method to support complex decisions. The model alone, while an essential foundation for an intelligent decisioning process, is only a part of the overall solution. There needs to be an application layer to provide the ability to interact with the model, visualize the information, add in needed automation / AI / ML, and track the state of information as the business endeavors to fill in the gaps for a complete model to support the decision process. The system could look something like this:
The application should provide:
Custom query / questions (scenario based and user configurable)
Data visualization to help resolve the “state” of data
Method to collect (or facilitate collection of) needed additional input data
Record and visualize progress
Add automation and AI to:
Continuously improve the model based on available data and outcomes from previous decisions
Provide a virtual assistant to interact with the data in a conversational way and assist in data collection and analysis
Add access to one or more language models (system design driven)
Integration to preferred collaboration tool(s)
Other considerations:
Integrated internal data - Internal data silos created by disparate data storage systems and stand-alone applications
There may be a need for real-time information streams based on decision type
For complex decisioning businesses need a complete information model and a system / application that enables the active use of the information including managing understanding around what information is available and what is needed to complete the model. Adding AI / ML and automation in the application provides a critical feature to continuously improve the model and the decisioning process. Using the system in a team / collaborative environment increases the accuracy over time, and can facilitate more effective employee onboarding / training and job satisfaction.
