LLM = ANALYSIS - PARALYSIS?
Guest Contributor: V. “Bala” Balasubramanian, PhD, MBA
In today’s AI (Artificial Intelligence) landscape, finance and business leaders rely on data more than ever to make strategic decisions and boost business performance. A large language model (LLM) like ChatGPT, CoPilot, and Gemini is an AI solution that can generate insights from data easily and quickly. In recent months, many enterprises are harnessing LLM tools for generating insights, and Gartner predicts that by 2026, over 80% of enterprises will have used LLM models [1].
However, the abundance of insights generated from these LLM tools has become a double-edged sword. With limited time and resources, leaders often find themselves overwhelmed by a flood of insights, leading to “analysis paralysis.” Analysis paralysis is a situation where excessive insights leads to overthinking, resulting in inaction. It often occurs when an individual or team is overwhelmed by too many insights resulting in a fear of making the wrong decision. According to research by Oracle and Seth Stephens-Davidowitz, 85% of business leaders have experienced decision stress, and three-quarters have seen the daily volume of decisions they need to make increase tenfold over the last three years. Poor decision making is estimated to cost firms on average at least 3% of profits, which for a $5 billion company amounts to a loss of around $150 million each year [2].
However, for most enterprises, generating insights from data and analytics was never a big hurdle—it was always about translating those insights into actionable business outcomes. As Nobel laureate Daniel Kahneman notes in his book “Thinking, Fast and Slow”, “more insights don’t necessarily lead to better decisions; they often complicate them.” [3] Just because, an LLM is generating insights, doesn’t mean productivity is guaranteed. In addition, validation of generated insights requires human in the loop, which is in short supply to begin with, potentially nullifying the effects of “so called” increased productivity.
So, what can enterprise do to manage the “analysis paralysis” problem in the world of LLM and derive business benefits? Here are three main steps to overcome the “analysis paralysis” problem and improve productivity in the enterprise.
1. Define the Decision Context or Objective. The key step in every insight derivation process is asking the right questions (known as prompts in LLM). Prompts or questions must be further bolstered by clearly articulating what decisions needs to be taken, what insights are required to make the decision, and what data is required to derive the insights. The context for decisions, insights, and data is business objectives.
2. Harness the various elements in the decision value chain. Insights from data can be derived manually by humans (i.e. analysts) or from LLMs. While deriving insights from LLMs has become significantly easier, turning those insights—whether generated manually or by LLMs—into actionable decisions that drive meaningful outcomes remains a major challenge for most enterprises. Decision making from insights goes beyond simply presenting data or KPIs or insights. It encompasses harnessing the right data, deriving insights, targeted messaging, creating trust and cultural awareness, addressing the ethics, bias, and compliance aspects, engaging with the right stakeholders, creating impactful visuals, demonstrating measurable business impact, and more. The process of making the decisions from insights in the LLM world is as shown below.
3. Use Decision Frameworks like Pareto Analysis, Pugh Matrix and Eisenhower Matrix.Pareto Analysis helps to identify the 20% of data and insights that drive 80% of the decisions. Pugh Matrix or Decision Matrix weighs options or alternatives against criteria like cost, time, risk, and potential impact to select the best alternative. Once the alternative is picked using the Pugh Matrix, there would be numerous tasks pertaining to achieving the alternative. Eisenhower Matrix separates those tasks into urgent vs important to prioritize those tasks effectively.
Today, LLMs have revolutionized data analysis, processing vast datasets to extract actionable insights with unprecedented speed. While these tools make uncovering insights easier and faster, the real challenge lies in translating those insights into tangible business results. LLMs, powerful as they are, cannot replace the essential human qualities of creativity, curiosity, and critical thinking [4]. Moreover, LLMs come with hidden costs. For example, in a mid-sized firm with 5,000 LLM users, the annual CO₂ emissions from their deployment (is about 60 metric tons) is comparable to those of 15 gasoline-powered cars—a significant environmental impact.
The “Human in the Loop” (HITL) emphasizes the need for human involvement at key stages of the decision-making process to ensure better outcomes, quality, and control. Businesses must resist the temptation to get lost in endless data, analysis, and even the hallucinations produced by LLMs. Instead, enterprises should focus on aligning insights with validation and measurable business outcomes. By bridging the gap between insights and action, organizations can overcome analysis paralysis and achieve meaningful progress, driving improved performance and lasting impact.
References
- https://www.gartner.com/en/newsroom/press-releases/2023-10-11-gartner-says-more-than-80-percent-of-enterprises-will-have-used-generative-ai-apis-or-deployed-generative-ai-enabled-applications-by-2026
- https://hbr.org/2023/10/how-ai-can-help-leaders-make-better-decisions-under-pressure
- Kahneman, Daniel; “Thinking, Fast and Slow”; Farrar, Straus and Giroux publishers; April 2013.
- https://www.forbes.com/councils/forbestechcouncil/2022/10/17/the-five-cs-soft-skills-that-every-data-analytics-professional-should-have/
Identify your path to CFO success by taking our CFO Readiness Assessmentᵀᴹ.
Become a Member today and get 30% off on-demand courses and tools!
For the most up to date and relevant accounting, finance, treasury and leadership headlines all in one place subscribe to The Balanced Digest.
Follow us on Linkedin!
