Using AI to Boost Senior Executive Performance
How You Should Use AI Tools to Improve Leadership and Decision-Making
Topic: AI, corporate management, decision making, personal productivity
Target Audience: All senior decision makers
Key Insight: AI tools can now be used to improve leadership and decision-making
Action Needed: Learn when and how to use AI tools to improve your work, and start using tools such as ChatGPT daily, to build a competitive edge
AI, Generative AI, and LLMs for Strategic Decision Making
Artificial Intelligence (AI) is no longer the future; it's here, making waves in the business world. In the last article, we saw it double the productivity of software developers, and now it's set to revolutionize the work of senior decision-makers.
In essence, AI enables machines to learn from their inputs, understand complex content, predict outcomes, and adapt to new information. This results in operational efficiency and more accurate decision-making – invaluable benefits for top-level management.
Generative AI, a subset of AI, goes a step beyond by creating unique outputs from the inputs it receives. This technology is particularly impactful in the field of natural language processing (NLP)
.
One of the most exciting developments in Generative AI is the advent of Large Language Models (LLMs). These models generate consistent, intricate text on a large scale, thanks to the extensive data they're trained on. As they grow, LLMs develop emergent abilities such as language translation, question answering, and content summarizing. For senior decision-makers, this represents an opportunity for substantial productivity boosts, enabling more strategic and informed decisions. So let's dive in and explore how embracing these AI tools can usher your operations into a more efficient, innovative future.
What is an AI Tool?
In this article, we're exploring 'AI tools', which are based on Large Language Model (LLM) technology. Before we proceed, let's clarify some definitions. AI tools are continually evolving, as are their classifications and business models. Here's how we're defining them in this context:
AI Engines: These are the powerhouses behind modern AI tools. Good examples include systems like GPT-4 or PaLM 2. These engines require an interface, such as a chatbot, for users to interact with them. Their true potential, however, is harnessed through their APIs, which enable other tools to leverage their capabilities.
AI Chatbots: These are AI-powered systems like ChatGPT and Bard. They run on one or more AI Engines and can be upgraded or modified without changing the chatbot's core functions or interface. When you use them, you input a prompt, and the engine generates a response.
AI-Enhanced Tools: These are applications built around an AI use case or traditional software like Office 365 or GitHub, which have been supercharged with AI engines like GPT-4. We're calling these enhanced tools 'Copilots', taking inspiration from Microsoft's terminology. When you use them, your interaction with the AI is indirect - you experience it through the improved functions the AI brings to the tool.
AI Agents: These are AI systems that combine multiple AI engines and other tools to perform specific tasks. They can turn basic instructions into complex actions, ranging from sending an email to writing a software application, or even planning a trip. Currently, we see these AI tools mainly in open-source experimental settings, like AutoGPT. When people talk about Artificial General Intelligence (AGI), they're envisioning an AI Agent that's capable of performing tasks as diverse and complex as a human can, across a broad spectrum of areas. AGI represents the next frontier in AI, promising systems that can understand, learn, and apply knowledge across a wide variety of tasks and adapt to new situations much like humans do.
Who are the Leaders in AI Technology?
As the realm of AI surges forward with incredible speed, a few pivotal players are advancing AI technology and applications. Here's a snapshot of these key innovators.
OpenAI is a standout in the arena, leading in the development of Large Language Models (LLMs). The crown jewel of their achievements is the AI engine GPT-4, which fuels their popular AI chatbot, ChatGPT.
Alphabet, the parent company of Google, has a rich AI lineage. Despite currently trailing in the LLM domain, its subsidiary DeepMind boasts numerous trailblazing AI discoveries. Alphabet is making significant strides with its AI engine PaLM 2, AI chatbot Bard, and Google Duet, their workplace copilot.
Microsoft is steadily gaining ground in the AI race, bolstered by its partnership with OpenAI. Employing OpenAI’s GPT models as its AI engine, Microsoft has introduced Bing Chat, an AI chatbot. They are also embedding AI into their main tools, like Office 365 Copilot and GitHub Copilot.
Although Meta, formerly Facebook, isn't leading in LLM development, it merits acknowledgment for open sourcing its models, thereby inviting wider community involvement. Their AI engine, LLaMA, forms the backbone of their AI initiatives.
Transitioning from tech titans to ambitious startups, let's consider some notable new players in the AI landscape. Anthropic, backed by a robust $1.3 billion funding and founded by ex-OpenAI members, has made waves with their chatbot, Claude. Stability AI, another startup to watch with $110 million in funding, is contributing to open-source AI with their StableLM model.
As we delve into this topic, our primary focus will be on ChatGPT, which is powered by OpenAI's GPT-4. Currently, it stands as an effective tool for a wide range of applications, which we'll investigate further here.
How Can AI Tools Empower Senior Decision Makers?
The motivation for integrating AI tools into the decision-making toolkit of senior leaders is twofold.
Firstly, these tools have the potential to significantly enhance performance. They can refine decision-making by offering data-driven insights, direct attention to critical issues by identifying trends and anomalies, improve communication clarity by automating routine correspondence, sharpen risk management by predicting potential issues based on historical data, and optimize time utilization through task automation. In today's fast-paced business environment, such enhancements are not just beneficial but increasingly necessary.
Secondly, gaining first-hand experience with AI tools equips senior executives with an understanding of this transformative technology. This isn't a fleeting trend; it's a powerful force set to bring about fundamental changes in the way organizations operate. By mastering these tools today, leaders can proactively prepare their organizations for the inevitable shifts of the future, instead of reacting when change becomes inevitable.
Can AI Tools Make Your Work Better Today?
AI tools are like your personal helpers, ready to lend a hand in all your tasks. Picture them as top-tier graduates, able to tackle a variety of topics. They're your tireless office sidekicks, ready to help you get more done, faster.
Condensing Information: AI is adept at distilling vast quantities of text into succinct summaries. This is equivalent to turning a hefty report into a short list of the main points. It's an efficient way to free up time for executives, allowing them to concentrate on making the big decisions.
Spotting What's Important: Another strength of AI tools is their ability to extract the most relevant information from a sea of data. They hone in on the key messages, thus providing decision-makers with just the essentials they need.
Help with Writing: AI programs, such as ChatGPT, can be a great help when crafting emails, reports, and proposals. They ensure not only grammatical correctness but also the effective communication of the intended message. Moreover, they can adjust the same message for different recipients based on length, complexity, and use of analogies.
Polishing Your Work: ChatGPT can also act as a rigorous editor, making sure your written work shines. Every piece of communication is honed to be professional and purposeful.
Advisory and Creativity: AI tools can offer advice, feedback, suggest best practices, and even come up with fresh, innovative ideas. They can sift through heaps of data, spot patterns, anticipate trends, and offer actionable insights, proving invaluable in strategic decision-making.
Prepping for Meetings: Prior to key meetings or talks, AI tools can conduct thorough research, giving you a firm grasp of new or complex topics. This helps ensure you're well-prepared, leading to more productive and informed discussions.
Risk Evaluation and Cross-Analysis: AI tools can spot potential risks and their consequences, conducting sophisticated cross-analysis of connected areas. By flagging potential issues and offering real-time alerts, they support proactive decision-making and risk mitigation. They can also review related initiatives and identify dependencies, overlaps, and contradictions.
The smart integration of AI tools into the daily work of senior executives and board members can enhance decision-making, improve communication, and boost overall performance. The key is knowing how to leverage these tools effectively to bolster executive capabilities.
Understanding the Limitations of AI Tools
AI tools, though incredibly powerful, come with their own set of challenges. A thoughtful approach to their use is necessary, particularly in the early stages. Here are some crucial aspects decision makers should take into account:
Data Security: AI tools weren't initially built with specific business applications in mind, opening up potential risks around data safety. The information fed to these tools could inadvertently influence future versions of the model, possibly leading to leaks of sensitive data. This emphasises the critical need for stringent data protection measures when employing AI tools.
Accuracy: AI tools may occasionally slip into "hallucination", creating false facts, inventing sources, or fabricating quotes. This could prove problematic in situations that demand absolute accuracy and reliability. Thus, it's always wise to cross-check the information generated by AI tools.
Skill Requirements: To use AI tools effectively, specific skills are needed. One such vital skill is 'prompt engineering' – the expertise of giving the right instructions to the AI to generate precise and beneficial results. Therefore, understanding the strengths and limitations of the AI tool is critical.
Model Capabilities: AI tools, like ChatGPT, have certain restrictions. For example, ChatGPT can only manage about 8,000 tokens, equivalent to roughly 6,000 words or 12 pages of written text. Any conversation beyond this limit will only consider the latest 8,000 tokens. Additionally, AI models have a cut-off date for their pre-learned information. For instance, GPT-4's cut-off is September 2021. Any information beyond this isn't pre-learned into the model, though it can access more recent data via a browser plugin - a feature still under refinement.
The AI field is progressing at breakneck speed, meaning the features, interfaces, and even business models tied to these tools can quickly evolve. The knowledge you gain about an AI tool today might not be applicable tomorrow, making ongoing learning and adaptability crucial for their successful use.
Example #1: Analyse Amazon’s “Letter to Shareholders” as if You Were a Board Member
Here is an example how you can analyse a long text, in this case the 10 pages Amazon 2022 letter to shareholders and extract key focus areas that you can use as a starting point for your own analysis.
Prompt: “You are now an experienced board member and skilled decision maker, known for focusing on the right things for the business. Analyse the letter: https://s2.q4cdn.com/299287126/files/doc_financials/2023/ar/2022-Shareholder-Letter.pdf and come up with the top five things a board member should focus on during the next year. For each of the five things, estimate how much of your time (in %) you should spend on the topic, and also estimate how large the risk associated with the topic is (0 to 100%)”
By asking ChatGPT to simulate a different role, you can enhance the analysis. Keep in mind, though, that this approach requires a ChatGPT equipped with a PDF handling plugin.
Example #2: Break Down the Complex for Easy Understanding
In a business world full of jargon and complex ideas, it's normal to come across technical terms. You're not required to understand every tiny detail, but having a basic grasp can prove helpful. Let's take "L2 zK-rollups" as a case in point.
Prompt: “Explain what an L2 zK rollup is. Explain it like I'm 12 years old. Add an analogy. Also explain a use case where a traditional company might use it.”
From the response, you can explore more, clarify doubts, or dive deeper into the subject matter.
Example #3: Transforming Rough Ideas into a Professional Email Proposal
AI tools have the ability to convert scattered thoughts into polished, cohesive documents.
Prompt: “Write an email to the management team proposing we launch a project to investigate potential of using AI/LLMs to improve productivity, based on my following unstructured notes: 3 weeks, cross company, led by HR, use internal resources, might be option to reach savings targets, not falling behind our competitors, use output as input in business planning, etc. Write the proposal in a professional, informal, convincing, and action-oriented style. Structure in the style of the Pyramid principle.”
The more information and context you provide to the AI, the better its output will be, minimizing the chances of producing nonsense or 'hallucinations'. Once the draft email is ready, the team members can review it, provide their feedback, and use their own AI tool to craft a revised proposal in a matter of minutes.
Reflections
Today's AI tools are highly advanced and hold immense potential to substantially elevate the performance of senior decision-makers. However, I anticipate that these individuals will not adopt AI as extensively as other groups, such as software developers, who are generally more accustomed to incorporating new technologies into their workflows. Senior decision-makers often prefer to rely on tried and tested strategies. Despite this, I firmly believe that AI represents a golden opportunity for these leaders to gain a competitive advantage.
That being said, to tap into the full potential of AI, several challenges need to be addressed and overcome:
Balancing AI Tool Usage and Data Security: One of the primary challenges is to find an optimal balance between robust use of AI and adherence to strict security standards. Given that decision-makers handle sensitive corporate data, the secure use of AI tools becomes paramount. While an outright ban on usage might seem the simplest solution, it carries a substantial risk of leaving the company at a disadvantage. For instance, some companies, like Samsung and Apple, have completely prohibited the use of generative AI on company-owned computers. This could lead to a competitive lag and the potential loss of talented individuals who could significantly boost their productivity with AI tools.
Developing Necessary AI Skills: Effective use of AI requires a unique set of skills, such as the ability to logically and critically direct AI systems – a process known as prompt engineering. Expert prompt engineers, commanding annual salaries of over $375,000, are in high demand. Failures often blamed on AI, including those associated with tools like ChatGPT, are often the result of imprecise inputs or questions. It's like blaming a typewriter for typing errors. For instance, instructing the AI to "write a poem" could yield very different results depending on whether the AI was previously given guidance on poem structures, best practices, thematic development, and context. Furthermore, more accurate results can often be achieved by asking ChatGPT for a step-by-step explanation using phrases such as, "Let’s walk through this step-by-step, to make sure there are no errors…", so called Chain of Thought prompting.
Finding New Ways of Working: Implementing AI in decision-making processes requires changes in work culture and the way organizations function. A notable example is Amazon, known for its document-based culture where narrative memos or “six-pagers” are used for decision making. Although this approach wasn't designed with AI in mind, it incidentally suits the use of language learning models (LLMs) perfectly. There are hence more reasons for other companies to copy similar ways of working now when it can impact AI efficacy.
Adapting to Rapid Change: AI technology evolves at an exponential pace. Executives and board members must be prepared to keep up with this rapid change to remain competitive.
The adoption and advancement of AI tools for improving the personal productivity of senior decision-makers is expected to follow a rapidly evolving trajectory. Used in the right way, AI tools can elevate the quality of output by providing data-driven insights and augmenting analytical capabilities, in the short term. In the long run, they offer the potential for time-saving automation of workflows, a shift that could transform the executive landscape. The evolution can be visualized as progressing through three primary stages:
Ad-hoc AI Usage (within 3 months): AI-assisted support in simple tasks. Prominent examples include AI assistants like ChatGPT, which can be used for a variety of tasks such as email drafting, scheduling, and more. Prompt engineering, which ensures immediate and accurate responses, is a key element at this stage. With the release of ChatGPT for Business, we anticipate broader adoption among senior executives and decision-makers.
AI-Workflow Integration (3-6 months): This phase marks the point where AI becomes an integral part of our daily routines, with AI-embedded tools blending seamlessly into operational workflows. The incorporation of AI into popular platforms like Office 365 will facilitate this transition. Such AI agents and tools enhance users' capabilities and oversight, serving as a valuable extension to their work.
Intelligent Organizational Transformation (12+ months): The final stage of this progression leads us to the emergence of intelligent organizations. Historically, organizations have been structured like Napoleon's army—decentralized, with clear hierarchies, specialized support units, and a coordinated strategy for unified operation. This model has been effective for over two centuries, but in the era of digital and AI transformation, we need a novel approach to organizational structure. To understand what this may look like we need to look in different directions. My guess is on the fusion of AI with ideas behind Decentralized Autonomous Organizations (DAOs), and agile methodologies. The result would be a dynamic, fast-moving organization with a decentralized, largely AI-automated framework, significantly enhancing productivity and decision-making capabilities.
Across these stages, AI tools enable senior decision-makers to not only simplify their individual tasks but also reshape their organizations holistically, nurturing a culture of adaptability, efficiency, and intelligent functioning.
Recommendations
Make sure you’re early to the game and get first-hand experience. Delve into the world of AI tools as soon as possible to gain first-hand experience. The value of these tools might seem limited now, but they are constantly improving. Each upgrade enhances their potential, so expect rapid growth in their capabilities and value over time.
Find a pragmatic balance between Data Security and AI tools usage. Understand your data landscape. Identify the data that must be protected at all costs and the data that, in the worst-case scenario, you could afford to reveal. Keep in mind that refraining from using AI tools entirely carries risks, particularly in terms of competitiveness and talent retention. If you can't stay competitive, the need to protect your data is irrelevant.
Focus on quick-win changes in ways of working to better leverage AI. Given the current capabilities of top AI tools, it's advantageous to have as much data as possible in a well-written text format. If you have ten pages of well-structured text, for example, you can condense it to any length or query it as needed. Today, this is more beneficial than dealing with an extensive PowerPoint presentation or a heavily redacted executive summary.
To conclude, AI tools has the potential to greatly support senior decision-makers. While there currently are challenges, their capabilities to improve decision making, risk management, communication exists already today, and it will improve in a fast pace. The sooner you can embrace it, the larger the potential to build an edge – and stay ahead of the curve.