Introduction
Today, we are inundated with more data than at any other time in history. Those in the field working to produce actionable intelligence assessments have the near-impossible task of sifting through mountains of data from myriad sources for one key insight. This insight is used by intelligence consumers, from tactical warfighters and first responders to strategic senior-level policymakers. It is imperative for our national security to meet this challenge head-on.
The solution hinges on the adoption and utilization of new technology like Artificial Intelligence (AI) to ease the burden of analysts, as well as collection and processing teams, to improve the intelligence production cycle by processing structured and unstructured data at scale and in near real time. We will explore how AI can add value to the national security landscape, as well as how Customer Success plays a role in these challenges.
The Current State
Intelligence gathering at the most basic level revolves around gathering valuable information to produce actionable decisions or give a leading advantage to an adversary. Today, tagging and indexing of information allows search functions to operate with relative ease, however, this is usually limited to highly structured and organized data. To create new knowledge and insight from an exploding universe of data, analysts must have the ability to link and analyze disparate data through structured data (databases) and unstructured data (news articles, annual reports, social media, etc.) at scale. This is an extremely labor-intensive task to determine people, places, and things in non-structured data, and is simply not feasible for teams to keep up with the constant flow of new information. Analysts have an uphill battle extracting actionable insights from the deluge of information.
The Value in Artificial Intelligence
Artificial Intelligence disciplines have been around for a long time, and the term was originally introduced in the 1950s. At a high level, AI encompasses technologies like natural language processing, machine learning, and predictive analytics to exercise self-learning problem-solving techniques. The real value of AI in the national security space is an end to organizational knowledge loss.
AI has the ability to capture the tacit domain knowledge of human experts and autonomously create semantic models or knowledge graphs that unify broad swaths of document collection and enable AI software called Agents to analyze, reason, predict, decide, and act at superhuman scale. Agents learn continuously from humans through natural interaction, learn from other agents via a multi-agent framework, build on accumulated knowledge, and have long-term memory.
AI Agents are superhuman digital twins of an organization’s best people that generate new knowledge and predictive insight long after the employee leaves the organization. AI Agents accrete or transfer accumulated knowledge across domains and generational cohorts, instantly connecting dots and producing predictive insights beyond human capacity. Likewise, AI can also break down language and cultural barriers by incorporating the nuances of human personality into natural interactions, all without any human in the loop beyond the user, that can provide feedback and quality assurance as needed.
Although Generative AI and large language, image, and voice models have soared in popularity, the rate of production deployment and retention in enterprise and government settings is abysmal. The reason for this is that absent a semantic layer or knowledge engine powered by tacit human domain knowledge, generative models alone cannot reason and produce insightful, predictive, and actionable content that’s useful to intelligence analysts in real-world mission-critical settings where explainability and source attribution in reports must comply with current analytic standards like ICD 203.
Customer Success and Change Management
Introducing AI tools into the mix isn't always a simple plug-and-play operation. Proper implementation requires good change management best practices, as well as an acknowledgment of a baseline operating model like the age-old People, Process, and Technology or a more modern approach with human-centered design. For most AI companies, implementation and customer experience fall somewhere within a Customer Success organization, especially in bridging the gaps of program management, change management, and implementation.
Proper implementation also requires visionary leadership that defines new ways of doing business as well as building trust with new AI technology. On the most basic level, it’s learning to work in the unclassified world and understand there is great value in open-source information. Scalability and compatibility are paramount, demanding flexibility with both internal and external challenges to your teams.
Closing Thoughts and Key Takeaways
So, instead of relying on old methods for information searches or hoping to create new knowledge and insight by infinitely promoting commodity generative AI models to no avail, let AI Agents powered by knowledge engines that capture and scale tacit knowledge that learn, reason, predict, decide, and act ease the burden of information overload and organizational knowledge loss. AI is not here to displace personnel but rather to augment and expand their capabilities and unleash human genius. Change is imperative to keep up with the insatiable demand for usable data and the ever-evolving capabilities of adversaries. Our role as Customer Success practitioners is to align with our clients’ goals and ensure our tools help achieve their objectives.
If you would like to learn more about this and similar topics, Accrete AI hosts a monthly networking event, Unsupervised Learning, at our office in Alexandria, Virginia. Contact me for more details!