Unleashing the Power of AI in Decision Support Systems
Decision-making is critical to the success of any organisation because it directly impacts the ability of a company to achieve its goals. According to McKinsey, over half of C-level executives spend over 30% of their time making decisions. This highlights the importance placed on decision-making at the highest levels of an organisation. However, the same amount of C-level executives believes that only 50% of their decision time is productive, which means that a significant portion of their time is being wasted on ineffective decision-making. This is not only a waste of time, but it also leads to poor decision-making and can ultimately harm the organisation.
Decision Support Systems in business
The market size of Decision Support Systems (DSS) varies depending on the source and the specific sub-segment of DSS being considered. However, overall, the global DSS market is expected to grow at a significant rate in the coming years.
According to a report by MarketsandMarkets, the global DSS market size is expected to grow from USD 11.6 billion in 2021 to USD 21.1 billion by 2026, at a CAGR of 12.3% during the forecast period. The market is driven by the increasing need for data-driven decision-making, the growing adoption of analytics and business intelligence tools, and the increasing use of big data and IoT in various industries.
Another report by Research and Markets estimates that the global DSS market size was valued at USD 12.5 billion in 2020 and is expected to expand at a compound annual growth rate (CAGR) of 11.9% from 2021 to 2028.
In terms of segment, the Advanced Analytics segment is expected to grow at the highest CAGR during the forecast period. The increasing need for advanced analytics to gain actionable insights and the growing adoption of big data analytics are some of the factors driving the growth of this segment. Another segment with rapid market expansion is the Military. My teammate shows an interesting point of view on AI in Modern Warfaremore and the role of Decision Support Systems.
What are decision support systems?
A Decision Support System (DSS) is a tool that helps people make decisions by providing them with relevant information, analytics, and modeling capabilities. DSS is used in various settings, such as company management, government, healthcare, and military operations. Here are a few examples from different industries:
- Personal finance — helps individuals make informed decisions about managing their money, such as budgeting and investment planning.
- Healthcare — helps doctors and other medical professionals make diagnoses, select treatments, and manage patient care.
- Retail — helps retailers analyse sales data and make decisions about inventory management, pricing, and promotions.
- Transportation — helps optimise routes and schedules for public transportation, trucking, and other types of transportation.
- Emergency management — helps manage resources and respond to emergencies and disasters.
One common misconception about Decision Support System is that it is a computer program. The early DSS relied entirely on humans, instead of depending entirely on technology. They relied on human expertise, rules, and heuristics to support decision-making. Such DSS can be as simple as a checklist or a flowchart that guides the decision-maker through a process.
One example of a human-powered DSS is an early disaster mitigation decision support system. This system is a set of guidelines, checklists and procedures that emergency responders follow when responding to natural disasters such as hurricanes, earthquakes, or floods. It helps emergency responders quickly identify potential hazards, prioritise responses, and make decisions about evacuation, rescue, and recovery efforts.
Building blocks of a DSS
A Decision Support System consists of three main component groups, which include the knowledge base, model, and user interface.
- Knowledge Base: The knowledge base is the foundation of the DSS and contains all the relevant data, information, and knowledge that the DSS needs to make decisions. The implementation of a modern knowledge base supports storing both structured and unstructured data. It includes data from a variety of sources such as databases, spreadsheets, data streams and external APIs. Additionally, model analysis, simulations and user feedback are a form of data and as such knowledge base stores them for future improvements of the system.
- Model: The model is part of the DSS that processes and analyses the data from the knowledge base. The model can be a simple rule-based system or a complex machine-learning algorithm. Outside of data analytics, models create simulations, identify trends, patterns, and insights, and make recommendations. Because of the need to process vast amounts of data, we use modern data architectures.
- User Interface: The user interface is part of the DSS that interacts with the user. It allows the user to access the knowledge base and model. The goal of DSS’s user interface is to allow users to fully focus on making decisions while providing all necessary data, simulation outputs and recommendations. This means that designing user-friendly and easy-to-use is one of the key points of focus for DSS developers. The user interface of a Decision Support System can be graphical, such as a web or mobile application. This is currently the most common option. However not every user has a huge screen to read all the charts and data. The newest trends in UX for AI models follow increased usage of voice-based chatbots and AR/VR-based user interfaces. Both of these options offer an interesting and convenient way to interact with DSS. Additionally, the producers of modern DSS invest in personalization based on user needs, preferences and tasks.
AI benefits
Processing of Unstructured Data
The use of Computer Vision (CV) and Natural Language Processing (NLP) in Decision Support Systems greatly improves their capabilities by allowing them to process and analyse huge amounts of unstructured data. CV helps with analysing visual data such as images and videos, while NLP focuses on analysing text data such as social media posts and customer reviews. This allows DSSs to gain insights from a wide variety of sources that would otherwise be difficult or impossible to process.
Self-improvement
The use of Reinforcement Learning (RL) in Decision Support Systems greatly improves their capabilities by allowing them to learn from experience and improve their decision-making over time. RL is a type of machine learning that focuses on training models to make decisions based on rewards or punishments. An example in the manufacturing industry is a smart factory, where DSS helps with production optimization processes. RL allows self-improvement based on past success metrics and feedback from the factory floor, which leads to better decision-making. Over time, the system learns from its experience and becomes more accurate and efficient. This leads to improvements in product quality, production cost and efficiency.
Multi-task Learning
Multi-task learning (MTL) models are beneficial for Decision Support Systems because they allow the models to learn from multiple tasks at the same time. Such models can generalise on similar or overlapping tasks, which improves system performance on new and unseen tasks. Especially when gathering data for each separate task is challenging or cost-ineffective, MTL models help to reuse already acquired and cleaned data. An additional benefit of MTL models is that they improve the robustness of DSSs by learning from multiple sources of data. In turn, it improves the accuracy of the decisions even if some data is missing or unreliable.
User Interface
Chatbots with Large Language Models in Decision Support Systems change the way users interact with systems around the world. Currently, using only graphical interfaces is far too limiting. Natural Language Processing (NLP) advances using natural language to ask about information with the context within people’s hands. Another example of change in the area of user interfaces is Virtual or Augmented Reality-based user experience. Generative Adversarial Networks (GANs) create life-like simulations which provide users with a more immersive and interactive experience. By fully leveraging these new ways of interaction, DSSs become more accessible and user-friendly.
Summary
Decision Support Systems powered by Artificial Intelligence are present in our daily lives. From healthcare and finance to transportation and retail, DSSs help organisations make better decisions. With the advances in AI, the capabilities of DSSs continually improve. Techniques such as Computer Vision, Natural Language Processing, Reinforcement Learning, and Transformers enhance the capabilities of DSSs. AI is revolutionising the design of DSS by allowing it to process more types of data, such as images and videos, draw important information from texts and provide novel methods of interacting with the DSS. An increasing number of companies adopt chatbots with Large Language Models as user interfaces, achieving smoother and more user-friendly interactions with the DSS.
However, the biggest challenge for humans is to use these powerful tools efficiently to our advantage. As DSSs continue to become more prevalent, it is important for individuals and organisations to understand how to use them effectively and ethically. This includes understanding the limitations of the systems and ensuring that the systems are transparent, explainable, and fair. By doing so, we make full use of DSS capabilities to make more informed and effective decisions and improve our lives and society.
For more MLOps Hands-on guides, tutorials and code examples, follow me on Medium and contact me via social media.
