In the dynamic realm of enterprise management, “Embracing Gen AI: Transformative Strategies for Crisis Navigation” presents a thought-provoking exploration into the integration of Generative AI (Gen AI) as a pivotal tool for navigating crises. Spearheaded by the insights of Ankit Rana who is a CTO at Polestar, this discussion traverses the evolution of crisis management from traditional methodologies to innovative, AI-driven approaches. Rana’s expertise covers a broad spectrum of analytics, including data management, predictive modeling, and statistical analysis.
Through a series of questions, the podcast uncovers the multifaceted role of Gen AI in enhancing decision-making processes, streamlining recruitment, and providing actionable insights from vast data landscapes. This conversation not only highlights the capabilities and potential of Gen AI in addressing contemporary challenges but also delves into the critical resources, organizational changes, and strategic patience required to harness Gen AI’s full potential in crisis navigation.
AIM Research: How do you define “crisis navigation” within the context of enterprise management, and what motivated you to discuss this topic?
“It begins with a project experiencing an escalation, necessitating the project team’s engagement in crisis management”
Ankit Rana: Crisis management is an integral part of our daily lives. When discussing crisis management, it’s important to clarify that we’re not solely referring to disaster recovery or business continuity planning. Instead, crisis management occurs daily. It begins with a project experiencing an escalation, necessitating the project team’s engagement in crisis management. It extends to situations like a client not receiving payment releases, which becomes a crisis for a salesperson aiming to collect payments and meet targets. Therefore, crisis management permeates every aspect of our daily lives, which is precisely why I chose to address this topic.
AIM Research: How has the transition from intuition-based to data-driven decision-making evolved in businesses, and what role does Generative AI play in this shift, particularly in crisis management?
“This shift from the traditional, hands-on coding approach to the current reliance on artificial intelligence for generating code is significant.”
Ankit Rana: My journey has experienced numerous paradigm shifts. Initially, when I engaged in coding, we relied on a trial-and-error approach due to limited assistance, necessitating extensive reading from numerous books. This phase was followed by what we refer to as the Google era, marking a shift from research and development to leveraging Google searches for coding solutions. We are now transitioning from reliance on Google searches to utilizing generative AI technologies. It’s become remarkably simple to request generative AI to write code, receiving immediate solutions.
This shift from the traditional, hands-on coding approach to the current reliance on artificial intelligence for generating code is significant. As we discuss crisis management, the importance of data-driven decisions becomes apparent. These decisions are not confined by an individual’s experience, offering organizations a robust tool for effective crisis management.
The term “Gen AI,” commonly known as Generative AI, signifies not just the technology but also the generation that is growing up with it. For this new generation, there is no paradigm shift in adopting generative AI as they are already accustomed to it. However, for the older generation, although this transition represents a significant change, it is recognized as beneficial. Decision-making processes are now based on facts and data rather than solely on personal experience.
AIM Research: What are some examples of Generative AI applications in crisis management, including the business contexts and outcomes?
“The limitation in the job market is not the number of jobs available but rather the availability of skilled talent.”
Ankit Rana: I will discuss one particular use case we are actively developing among many. This use case focuses on recruitment, a fundamental need for every company. Traditionally, recruitment involves filling out numerous forms, screening countless resumes, conducting interviews, and then finally offering a job. This process, while straightforward, is time-consuming and inefficient for many organizations.
Our approach simplifies this process through the implementation of a bot, powered by generative AI. For instance, if a company wants to hire an individual with five years of experience in a specific technology, the bot can automatically generate a job description. It then prompts the user to review and approve this description or make necessary adjustments. Upon approval, the bot disseminates notifications to relevant parties, including recruitment managers and HR teams.
Furthermore, it integrates with job portals’ APIs to publish the job opening. Beyond posting the job, the bot searches the company’s resume database, ranks candidates based on relevance to the job description, and filters the most suitable resumes. Selected candidates are then contacted by the bot to schedule interviews, which are organized through various platforms provided by Original Equipment Manufacturers (OEMs).
During the interview phase, the generative AI offers assistance by suggesting questions of varying difficulty levels (easy, medium, or complex) to the interviewer. Moreover, using AI co-pilot tools, interviews can be recorded for further analysis, such as evaluating the candidate’s expressions and ensuring their speech is in sync with their lip movements. This capability to analyze video interviews adds a layer of depth to the recruitment process, enhancing the decision-making process.
This recruitment use case demonstrates a significant generative AI application, streamlining the hiring process and improving efficiency and effectiveness. It’s a prime example of the innovative solutions we’re working on and has been well-received for its practical value.
Talent management, or more specifically talent acquisition, represents a significant challenge for everyone. At Polestar, for instance, we conduct interviews throughout the year, often scheduling them for nearly every other Saturday or Sunday. This necessitates having interviewers available to conduct these interviews, a situation that is generally met with reluctance. The prospect of working on a Saturday might be acceptable once, but the idea of doing it repeatedly is understandably less appealing. This scenario underscores the ongoing need for crisis management in talent acquisition. As I often point out, the limitation in the job market is not the number of jobs available but rather the availability of skilled talent. Consequently, every organization must engage in crisis management concerning talent. This is the rationale behind our continuous efforts in this area.
AIM Research: Where do you find Generative AI’s role in crisis management?
“Turning to the application of GenAI in crisis management, consider the vast amounts of data every organization possesses, often likened to oil for its value.”
Ankit Rana: I’ll address this question from two perspectives. Initially, regarding Generative AI (GenAI) and its current stage of adoption, we find ourselves at an early, nascent phase. Generally, any technological adoption follows three stages: the defense stage, the extended stage, and the appending stage, where the technology is utilized to its fullest potential. While artificial intelligence has been around for over a decade, the advent of generative AI marks a new, embryonic phase. Currently, our use of generative AI is primarily focused on advanced search, summarization, and code generation. However, in terms of its generative capabilities, we still have a significant journey ahead.
Turning to the application of GenAI in crisis management, consider the vast amounts of data every organization possesses, often likened to oil for its value. Yet, the challenge remains not just in collecting data but in extracting actionable insights from it. Typically, data is presented in charts and reports, with root cause analysis requiring extensive effort. Enter generative AI, which revolutionizes our ability to query data, asking any question in natural language. This technology introduces contextualization, a previously missing component in AI, significantly benefiting crisis management.
Regarding the second point, the necessity of human involvement in crisis management remains paramount. Crises involve humans, not systems, and while GenAI can provide solutions, it’s the human element that navigates these crises. The process involves identifying a crisis, consulting GenAI for potential solutions, and then, crucially, innovating and applying these solutions effectively. This human-driven innovation in response to GenAI’s suggestions represents a major shift in managing crises.
AIM Research: In the context of utilizing Generative AI for crisis management, beyond the indispensable role of data, what additional resources are necessary to ensure its effectiveness in addressing enterprise-level problems and crises, particularly emphasizing the importance of organizational change?
“Beyond the initial understanding of Gen AI, the integration requires substantial modifications in business workflows, processes, and roles”
Ankit Rana: The first change that organizations and leadership must undertake revolves around the understanding of Generative AI (Gen AI). There’s a common misconception that Gen AI is simply akin to a chatbot, limited to answering questions and drafting emails. However, its potential for learning and evolution is significantly greater. Recently, I discussed with a customer their concern about naming conventions within their organization and how to train Gen AI to recognize different terms for the same concept. This training process is analogous to how new employees are introduced to company-specific terminology; Gen AI learns in a similar manner.
Beyond the initial understanding of Gen AI, the integration requires substantial modifications in business workflows, processes, and roles. This adjustment is crucial from a business perspective. Additionally, from a talent management standpoint, assembling an implementation team, providing necessary training, and most importantly, securing the patience and support from leadership for this team are vital components.
On the technical front, the adoption of Gen AI necessitates consideration of cybersecurity, data privacy, and the mitigation of bias and fairness. These elements are crucial for responsible creation and deployment of technology. These are the primary resources and considerations that I believe are most critical for effectively integrating and leveraging Generative AI within an organization.
AIM Research: What does requesting patience from your end customers entail as you develop Generative AI applications for crisis management, and how does this patience contribute to enhancing their work processes and improving crisis response?
“Experiences with chat GPT have shown that asking the same question twice can yield different responses. This variability highlights the importance of patience and the value of providing feedback, whether to affirm the relevance of an answer or to indicate its irrelevance”.
Ankit Rana: Patience encompasses several aspects, especially when integrating Generative AI (Gen AI) into operations. Initially, it’s crucial to set realistic expectations: Gen AI is not infallible and will evolve over time. This understanding is vital from the customer’s perspective, emphasizing the need for a feedback loop whenever a query is posed. For instance, experiences with chat GPT have shown that asking the same question twice can yield different responses. This variability highlights the importance of patience and the value of providing feedback, whether to affirm the relevance of an answer or to indicate its irrelevance. Such feedback loops are essential for Gen AI to learn and improve.
Implementation might seem straightforward, but the underlying neural science, focused on learning and adaptation, is more complex. Another key area requiring patience is ensuring that Gen AI initiatives align with strategic goals rather than being pursued as mere novelties. It’s important to differentiate between operational and strategic objectives; the latter typically takes longer to achieve due to their broader impact and significance. We advise our customers to understand and embrace this distinction, recognizing that strategic goals necessitate patience and a long-term view for successful integration and realization of Gen AI’s potential. This strategic alignment is what we emphasize for achieving substantial, lasting value with Gen AI implementations.
AIM Research: Considering the uncertainty in technology trends and the significant patience required for Gen AI development, how do you plan to manage if Gen AI fails to meet expectations for crisis management, potentially leading back to a reliance on human intelligence?
“Gen AI’s value is maximized when it’s employed towards enhancing process optimization and achieving strategic objectives, thereby yielding tangible benefits.”
Ankit Rana: I doubt that particular thing. The transition from one technology to another is, indeed, a constant; as the saying goes, “change is the only constant.” You can’t halt progress. It’s accurate that within this decade alone, we’ve witnessed the rapid rise and subsequent decline of several technologies. For example, big data and augmented reality/virtual reality (AR/VR) were hailed as groundbreaking, only to see their hype eventually deflate. Specifically, the big data movement was once synonymous with Hadoop. Today, while some organizations still utilize Hadoop’s architecture, it’s largely been superseded by more advanced cloud technologies, leading to the metaphorical “bursting” of its balloon.
Regarding AR/VR, its initial allure was primarily based on its novelty rather than strategic utility. This underscores my belief that technology must align with strategic goals to be genuinely effective. If organizations view Generative AI (Gen AI) merely as a tool for composing emails or generating marketing content, they’re unlikely to realize its full potential. Gen AI’s value is maximized when it’s employed towards enhancing process optimization and achieving strategic objectives, thereby yielding tangible benefits.
The question of whether Gen AI might become obsolete also merits consideration. Given the pace of innovation observed in this decade, it’s conceivable that what seems indispensable today might be replaced tomorrow by an even more advanced technology. So, while there’s a possibility that the current iteration of Gen AI could fade into obsolescence, it’s likely to pave the way for a next-generation AI, continuing the cycle of technological evolution.
AIM Research: In the foreseeable future, do you predict that AI systems such as Gen AI will predominantly aid companies in resolving critical dilemmas and crises, or do you envisage a scenario where AI could entirely replace human decision-making, potentially leading to AI being a CEO?
“I propose it will be EAs to the CEOs, not the CEOs themselves.”
AIM: To address that specific question succinctly, it wouldn’t be a CEO but rather an EA to the CEO. That’s where I envision this concept. When I mention generating AI, initially, I believe you mentioned that any type of AI operates akin to a system, correct? Now, considering a computer from a certain perspective, there’s a saying that a computer is only as intelligent as its user. For instance, if you input a simple program in any language, whether it’s C, C++, Java, or any other, and instead of “hello world,” you type “you are stupid,” the system will reflect back “you are stupid.”
Hence, even in Generative AI, it’s primarily about how you frame a question. In our experience with Generative AI, we’ve noticed a new role emerging known as prompt engineering. Without prompt engineering, merely posing questions will yield generalized responses. This is where human intervention becomes necessary. As I mentioned, the forthcoming process entails consulting Generative AI for an answer, innovating upon that response, and then implementing it. That’s where the value lies. Therefore, CEOs and CEOs will continue to be human roles. Engineering AI will serve as an assistant. For instance, if a decision needs to be made, the AI can present several options, from which a choice must be made. This is why I propose it will be EAs to the CEOs, not the CEOs themselves.
