FORESIGHT: TRANSFORMING DECISION MAKING PROCESSES USING AI
Decision Analysis Services (DAS) is an independent professional services company. We provide professional services globally to clients within the Energy, Defence, Government, Health, and Transport sectors. We specialise in bridging the gap between strategy and operational execution.
DAS contributes to research, via our sponsored PhD and MSc students, using AI and ML tools to interrogate data and provide foresight to decision makers. This, combined with the work undertaken by our Innovation Lab places us in the position of being able to capitalise on, and incorporate AI, ML and Data Science developments in work for our clients.
INTRODUCTION
In today's fast-paced digital world, the development and deployment of digital processes has transformed the way businesses operate. While these processes generate vast amounts of data, their potential extends far beyond their initial purposes. Data science approaches such as the construction of comprehensive system models, predictive analytics, and text analysis can be used to unlock the potential of organisation’s data.
Decision support systems are fundamental tools for organisations. These systems encompass diverse technologies and procedures aimed at providing insights and recommendations to facilitate decision-making. This is typically achieved by distilling extensive data sets into accessible formats, highlighting crucial trends and patterns which are presented though interactive dashboards and reports. Many decision support systems, while proficient in data presentation and analysis, often lack proactive and predictive functionalities. This shortfall emphasises the need for innovative technologies capable of fully leveraging organisational data.
In this article, we discuss how data science approaches such as Artificial Intelligence (AI) can utilise the data, made available by various digital processes, to revolutionise decision-support systems.
BACKGROUND
Recently, there has been a surge of methods to enhance the use of data for decision making, including dashboarding tools such as Power BI and Qlik. These methods have demonstrated the value of data analysis by providing visuals, statistics, and key performance indicators to provide insight into the data and, therefore, the system it underpins. Despite this, these tools are limited in scope, providing data analysis functions but lacking the proactive and predictive decision support capabilities that organisations are increasingly demanding.
Innovative technologies, such as artificial intelligence (AI), machine learning, and data science, are emerging as a game-changer for decision support systems. These methods can uncover hidden patterns, trends, and correlations, whilst also adapting and evolving over time having learned from new data. By harnessing their power, decision support systems are set to revolutionise the use of data to anticipate future trends, mitigate risks, and capitalise on emerging opportunities.
SUPPORTING DECISION MAKING USING AI
Machine Learning (ML) is a form of AI that uses data and algorithms to enable AI systems to mimic the way humans learn, progressively improving over time. ML has the potential to have a significant impact on many decision support systems. For example, signal processing.
Signal processing involves converting and transforming data in a way that allows us to see patterns, trends and anomalies which are infeasible to analyse via manual inspection. Current approaches to analysing signal data, such as time-series sensor data, only provide summary statistics and visualisation, which is limiting. However, machine learning techniques can enhance this analysis, by examining historical data to make predictions on future data. It offers capabilities such as forecasting future predictions, detecting trends, and identifying hidden patterns that are causing the area of focus. These techniques could, for example, predict faults in engineering systems and support operators to optimise maintenance schedules. These increased insights enable a far more comprehensive view of the system. Which in turn, facilitates better decision making.
Further to this, when signal processing data is collect by systems, anomaly detection techniques can identify patterns or events that deviate from normal behaviour. This can be a supervised process (learning from a set of previously identified anomalies) or unsupervised (establishing normal behaviours from a wealth of data and thereby detecting outliers). Such an approach could be highly beneficial for detecting rare adverse events, for instance, in energy production, and in identifying health markers for clinical screenings, to name only a few.
Unstructured text is another rich source of data; however, its complex nature can make analysis challenging. A developing machine learning technique called Natural Language Processing (NLP) could help. NLP enables machines to understand, interpret, and generate human language, such as text. Its capabilities include summarisation, entity extraction and similarity search. These skills could provide powerful, analysis-ready insights to be integrated into existing decision support platforms. As a result, NLP could significantly enhance decision-making in processes where extensive text data is collected, such as risk management, safety case management and incident reporting.
Python, the most widely used programming language by data scientists, can be incorporated into many decision support platforms via APIs, or by python scripts being applied to data before it is ingested by dashboarding tools. This means that many of the innovative techniques discussed in this article can be easily integrated into existing decision support tools, significantly enhancing them without the considerable investment of a brand-new tool.
CONCLUSION
The pace of AI, machine learning and data science development is revolutionising decision support tools and platforms, significantly enhancing their capabilities and their impact. Promisingly, newly emerging techniques could take this even further. For instance, developments in language models could change the way we interact with decision support tools altogether. Rather than having to understand how to use complex tools to generate insights, they make it possible to simply request a particular analysis. Moreover, multi agent systems, which combine various AI tools, could allow for ever more complex analysis to be performed based on simple user inputs.
The future of decision support is poised for substantial integration and enhancement of capabilities through the transformative AI and machine learning advancements discussed in this article. This trajectory promises to revolutionise how decisions are made, empowering organisations to significantly improve the efficiency and performance of their systems and processes.