The intelligence cycle [fig 1], comprising collection, processing, analysis and dissemination, has traditionally relied on human analysts to interpret and act upon information. Recent advancements in artificial intelligence (AI) have the potential to augment human capabilities by automating routine tasks and enhancing decision-making processes.
This paper examines the intersection of AI and the intelligence cycle through a review of emerging techniques and how they may help with intelligence analytics. We identify five key use cases for AI within the intelligence cycle:
We suggest that AI can significantly improve the efficiency, accuracy and timeliness of intelligence operations. However, we also highlight the need for careful consideration of data quality, bias, provenance and explainability in AI-driven intelligence processes. Ultimately, this paper aims to provide a review to extend the understanding of the role of AI within the intelligence community, informing the development of more effective and efficient human-AI collaboration models.
Figure 1:
By introducing AI and automation into the Intelligence Cycle, current systems are starting to evolve into Decision Intelligence (DI) applications. DI refers to a class of technologies that enable organisations to make better decisions by providing insights, recommendations and automated decision-making capabilities. It involves the integration of data, analytics and machine learning algorithms to support informed decision-making.
A Decision Intelligence platform can provide additional capabilities to intelligence analytics by automating complex data analysis, providing actionable insights, and enabling analysts to focus on higher-level tasks, such as strategic decision-making and predictive modelling. By integrating AI and visual analytics tools such as i2, a DI platform can help analysts to quickly identify patterns and trends in large datasets more efficiently, leading to more informed decisions and situational awareness.
So, let’s now consider those five key use cases for AI within the intelligence cycle:
Signal detection in intelligence analysis is like finding the needle in a haystack, or sometimes the needle in the needle stack. Analysts use tools, techniques and procedures together with specialist tradecraft to identify meaningful patterns in data, distinguishing them from background noise. This skill helps detect potential threats, gather valuable intelligence and stay ahead of adversaries.
Signal detection plays a pivotal role in intelligence operations, involving the identification of potential threats or significant events that may impact decision-making processes. Today, where data volumes are vast and complex, Predictive Analytics is enhanced as a transformative tool, leveraging advanced machine learning models to enhance signal detection efficiency and accuracy.
Predictive analytics harnesses techniques, such as Decision Trees, and Neural Networks, to analyse historical and current data patterns. These models perform well at identifying subtle trends that might go unnoticed by a human analyst. AI when applied to predictive analytics can help select the best models for a given problem, reducing the risk of choosing a suboptimal model that may not perform well on new, unseen data. This is particularly important in situations with many interacting factors, where traditional statistical methods may struggle to capture complex relationships between variables.
AI also enables automated feature engineering, which reduces the need for manual feature engineering and improves the efficiency of the predictive modelling process. AI algorithms can handle high-dimensional data more efficiently than traditional statistical methods, allowing for more accurate predictions in situations with many variables.
AI predictive analytics systems can and should continuously monitor data streams, too, and adapt to changing conditions in real-time, providing a competitive advantage in many application areas. Additionally, these systems can provide insights into how their predictions were made, enabling data scientists to better understand the relationships between variables and improve model performance. For instance, in financial intelligence, AI-enhanced predictive analytics can more quickly flag suspicious transactions indicative of fraud or money laundering. Similarly, in Social Media monitoring, predictive analytics can detect anomalous behaviour that may signal impending threats. This processing of large data sets quicker enables faster and more informed responses; crucial capability in dynamic intelligence environments where timely action can significantly impact outcomes.
Machine learning (ML) can play a critical role in enhancing the effectiveness of intelligence analysis by automating tasks, improving accuracy and accelerating decision-making. By applying ML algorithms to language processing tasks such as text classification, sentiment analysis, named entity recognition (NER) and machine translation, analysts can quickly identify relevant signals and extract valuable information from unstructured data.
One significant way that ML is transforming the intelligence cycle is through the automation of initial screening. Rather than a user manually reviewing vast amounts of text for relevance, ML tools can quickly scan data to identify potential leads worth further investigation. This frees up analysts to focus on more in-depth analysis of those leads, improving the overall efficiency and effectiveness of the workflow.
ML can potentially significantly improve accuracy in intelligence analysis. By helping to reduce errors and inconsistencies associated with manual analysis, ML supports the analyst to ensure that critical information is not missed or misinterpreted. An important consideration when using ML is the issue of data-bias and AI ethics and the need to track bias in both data and models. Additionally, ML can provide real-time insights into emerging trends and patterns, enabling analysts to stay ahead of rapidly evolving situations.
The potential benefits of ML for language processing in intelligence analysis are broad, with examples ranging from social media monitoring to open-source intelligence (OSINT). For instance, ML tools can quickly analyse social media interactions to identify potential threats or shifts in public opinion. Similarly, analysts can use ML to automate the extraction of valuable information from publicly available data sources.
Automation by integrating machine learning for language processing is essential for intelligence analysts who need to stay ahead of emerging threats and capitalize on the strategic advantages offered by this technology. By doing so, they can augment their capabilities, reduce manual errors and enhance the overall effectiveness of the intelligence cycle.
In the realm of object detection, computer vision techniques, where computers can be made to gain high-level understanding from digital images and videos, have emerged as a powerful tool for analysts to identify and track threats and intelligence.
Object detection using computer vision enables analysts to identify objects of interest, such as vehicles, buildings, individuals or events, by applying machine learning models to large datasets. This technology allows for faster and more accurate analysis than traditional manual methods, reducing the risk of human error and increasing precision.
Computer vision also facilitates the tracking of movement patterns, providing analysts with a comprehensive understanding of the behaviour of the subjects in the investigation. Additionally, the use of computer vision enables analysts to analyse environmental context, including terrain, weather conditions and time-of-day, which is crucial for situational awareness.
The integration of computer vision for object detection into the intelligence cycle offers several benefits for analysts. By leveraging this technology, analysts can improve their accuracy, increase their speed and enhance their overall understanding of the threat environment. This, in turn, enables more informed decision-making and response times, ultimately supporting national security objectives.
To overcome these challenges, it is essential to address the limitations of current technologies and develop new methods for collecting, analysing and disseminating object detection data. By leveraging computer vision for object detection, analysts can augment their skills and enhance the overall effectiveness of the intelligence cycle, ultimately supporting informed decision-making.
Computer vision algorithms can analyze data from different sources, such as satellite images, aerial photos, or video from security cameras, drones, and body cameras. They help improve our understanding for intelligence or investigations.
The integration of computer vision for object detection into the intelligence cycle offers significant benefits for analysts, including enhanced accuracy, increased speed and improved situational awareness. By addressing the challenges associated with this technology and developing new methods for collecting, analysing and disseminating object detection data, we can unlock new opportunities for informed decision-making and more broader objectives such as investigations and strategic analysis.
The adoption of Natural Language Generation (NLG) into the intelligence cycle can have a profound impact on the way intelligence information is produced and shared. By automating report writing and analysis, NLG allows intelligence analysts to focus on higher-level tasks.
One of the most significant benefits of NLG is its ability to improve reporting efficiency and accuracy. Traditional reports can be time-consuming to produce and are prone to errors, while NLG generated reports are concise and consistently formatted. This enables analysts to provide more accurate and timely information to stakeholders, which is critical in today's fast-paced intelligence environment.
NLG can facilitate the creation of real-time dashboards that provide a comprehensive view of complex data sets. By integrating NLG with visual analysis tools, organisations can create interactive dashboards that showcase key indicators, trends and insights in real-time. This enables decision-makers to quickly assess emerging threats, identify areas of concern and make informed decisions for either operational or strategic outcomes.
The use of real-time reports as dashboards also enables organisations to communicate more effectively with a broader audience. By providing visualisations and narratives that illustrate complex data sets, NLG-generated dashboards can be used to inform strategic plans, allocate resources and engage with policymakers and other stakeholders. This is particularly important in today's era of rapid information sharing and collaboration.
The integration of Natural Language Generation into the intelligence cycle has the potential to revolutionise the way intelligence information is produced, analysed and disseminated. By automating report writing and analysis, creating real-time dashboards, and facilitating data sharing and collaboration, NLG can help organisations to provide more accurate, timely and effective information to stakeholders.
As intelligence analysis continues to evolve with more complex datasets with more volume, there is a growing recognition of the need for more effective and efficient decision-making processes. This area concentrates on the feedback of the intelligence cycle. The intelligence cycle is composed of a series of steps that keep repeating. The final step focuses on the overall goals and processes to ensure continuous improvement.
In recent years, AI has emerged as a key enabler of improved decision monitoring and auditing in intelligence analysis, enabling analysts to enhance their ability to monitor and evaluate the effectiveness of their decisions, thereby improving overall outcomes. One way this is achieved is by using automated review and analysis tools, which can quickly review large volumes of data, identify patterns and flag anomalies that may have been missed by human analysts.
Machine learning algorithms can be trained on historical data to predict future trends and outcomes, enabling analysts to anticipate potential consequences of their decisions. This predictive capability is particularly valuable in assessing the likelihood and impact of different scenarios, allowing analysts to make more informed decisions about resource allocation and prioritisation. Furthermore, AI-powered tools can provide risk assessments that are grounded in empirical evidence, rather than relying on intuition or anecdotal experience.
The role of AI in auditing is equally important. By leveraging AI-powered tools, organisations can detect inconsistencies and errors in data with greater ease, ensuring that information is accurate and reliable. This is particularly critical in the context of compliance monitoring, where AI algorithms can be used to scan large datasets for non-compliance or breaches of regulations.
By providing a clear record of decisions and actions, AI-driven audit trails promote greater transparency and accountability in intelligence analysis. Ultimately, the integration of AI into the Intelligence Cycle enables analysts to make more informed decisions, with greater accuracy and efficiency.
Using AI for decision monitoring and auditing can potentially offer great benefits to organizations. They can achieve better accuracy, work more efficiently, and be more transparent. These advantages are critical in the field of intelligence analysis, where the stakes are high, and the consequences of error or inaction can be severe.
The adoption of Artificial Intelligence in the intelligence cycle has the potential to significantly enhance human intelligence capabilities. However, it is crucial to address some important challenges such as the availability and quality of data, and the fusion of data. This can all affect the accuracy and reliability of AI-influenced intelligence and are core to DI.
Augmenting human intelligence with Explainable AI (XAI) and transparent decision-making processes is essential to build trust in AI-driven insights. Incorporating visual analytics into the AI-powered intelligence cycle enables humans to comprehend and understand complex data-driven intelligence, facilitating informed decision-making. By prioritising these critical factors, Decision Intelligence Platforms have the potential to unlock the full potential of AI in enhancing human intelligence, ultimately supporting more effective and efficient decision-making processes.
i2 acknowledges that analysis is a multifaceted skill, defying straightforward automation. While AI excels in processing vast datasets and identifying patterns, critical thinking - the cornerstone of human analysis - remains uniquely human. Augmented intelligence tools, such as i2, can facilitate insights and streamline tasks, but they are just assistance to a wider process. Human analysts must still interpret results, contextualise findings, and make informed decisions. By acknowledging these limitations, we can harness AI's strengths while preserving the nuances of human intelligence, ensuring that analysis remains an intuitive, iterative, human-driven activity.
The i2 Research & Development team is actively engaged in AI and ML projects to ensure that i2 has customer needs and requirements at the forefront.
About the author - Adam Etches
Adam is a technologist with over 25 years’ experience in helping customers understand and adapt new and emerging technology. Having started his career in academic research his specialism is focused on figuring out novel ways to discover and analyse interesting and challenging data. He is currently Director of Technology at i2 Group.
© 2025 i2 Group / N. Harris Computer Corporation. All trademarks owned by N. Harris Computer Corporation.
The Bradfield Centre, 184 Cambridge Science Park, Milton Road, Cambridge, CB4 0GA, UK
