Generative AI in the Enterprise: Asking the Only Three Questions That Still Count

We just finished reading Ken Fisher ’s book The Only Three Questions That Still Count: Investing by Knowing What Others Don’t.? Investing professionals and stock market enthusiasts may be scratching their heads right now, asking what Ken Fisher and artificial intelligence have in common.?This simple answer is absolutely nothing.?So, who is Ken Fisher and why is he being referenced in an article about #AI – and 12 years after this book was published in 2012?

Ken Fisher is a renowned investor, author, and founder of Fisher Investments , a global money management firm. He is widely regarded in the finance industry for his innovative approaches to investing and market analysis. Fisher's contributions to financial literature include several bestselling books and a long-running Forbes column, where he provided insightful and often contrarian perspectives on market trends and investment strategies. His profound impact on finance is underscored by his development of the Price-to-Sales (P/S) ratio, a widely used financial metric. Ken Fisher's ability to challenge conventional wisdom and his extensive expertise have solidified his reputation as a leading figure in the world of finance.

A ha!?This book inspired the Intelligent Enterprise Leaders Alliance because of its key theme…challenging conventional wisdom.?In other ways, Ken liked to zig when others where zagging.? It made him a fortune.?While we aren’t out to apply these principles to investing in AI-related equities, we are keenly intrigued about applying the questions to the current widely-accepted AI landscape and thought you might be too.

The Three Questions

The crux of Fisher’s work aims to provide investors with a framework to outperform the stock market by addressing three key questions.? Again, the aim of this article isn’t to help you beat the market – although that would be nice.? These questions, as applied to #GenerativeAI in the enterprise, are meant to get you thinking and perhaps planning for the unexpected (positive and negative).

Without further ado, they are:

1. What do you believe that is actually false?

Fisher emphasizes the importance of challenging widely held beliefs and assumptions in the market. By identifying and understanding what the general market believes that may be incorrect, investors can gain a contrarian advantage.?

Here’s a basic example:?Conventional investment wisdom suggests that debt?(country or company) has an adverse impact on stocks. Not true.?You’ll have to read the book for the proof ??

2. What can you fathom that others find unfathomable?

This question encourages investors to think deeply and creatively about the future and to identify opportunities that others might overlook or dismiss as impossible. By doing so, they can uncover unique investment opportunities.?

For example, imagine a scenario where the stock market has experienced a significant downturn, and the general sentiment is overwhelmingly negative. Most investors are selling off their stocks, fearing further losses. Fisher would argue that, in such situations, many stocks are likely undervalued because the panic-driven selling doesn't reflect the true long-term value of these companies.?

By fathoming that the market will eventually recover and recognizing that the underlying fundamentals of many companies remain strong, an investor can start buying high-quality stocks at a discount. In essence, while others see only doom and gloom, an astute investor who understands market cycles and investor psychology can fathom the potential for recovery and invest accordingly.?The sad fact is that most investors buy and sell and the wrong times, severely impacting long terms returns.

?3. What the heck is my brain doing to mislead me now?

Fisher highlights the significance of recognizing and overcoming cognitive biases that can negatively impact investment decisions. By understanding these biases and how they influence behavior, investors can make more rational and informed choices. One simple example involves the concept of Recency Bias, which is the tendency to give undue weight to recent events when making decisions.?

Imagine an investor who has recently experienced a sharp decline in the stock market.

Due to recency bias, this investor might overestimate the likelihood of continued declines and make overly conservative investment choices, such as moving entirely into cash or bonds, fearing further losses. This reaction is driven by the emotional impact of the recent downturn, rather than a rational assessment of market fundamentals and long-term trends.

Now that we’ve covered the questions, let’s take a stab at hypothesizing what they might mean for the future of Generative AI in the enterprise.

1. What Do You Believe About AI That Is Actually False?

Generative AI is often being touted as the magic bullet that will solve virtually all business problems. However, this belief might be (probably IS) dangerously oversimplified. The reality is that while Generative AI can – and likely will - perform astonishing feats, it is not a panacea.

Myth #1: Generative AI is infallible.

Enterprises must recognize that Generative AI systems are only as good as the data they are trained on. Biased data leads to biased outcomes, and no amount of processing power can overcome this fundamental issue. Believing that AI will always provide the correct answer WILL lead to costly mistakes and ethical pitfalls.

Consider the issue of AI Hallucinations—instances where AI models generate information that is entirely fabricated yet presented as factual. This is particularly concerning in applications like customer service, where an AI might confidently provide incorrect information, leading to customer dissatisfaction and potential legal issues. In more sensitive areas like healthcare, an AI's incorrect diagnosis based on flawed data can have dire consequences.?

These are pretty basic examples, but we think you get the gist.?

Start working through all these what-ifs and you’ll easily see that AI can – and may – actually create a lot of rework, unwinding and damage control.

Another pitfall is Algorithmic Bias, which arises when AI models trained on biased datasets perpetuate and even amplify these biases. For example, if an AI recruitment tool is trained on historical hiring data that reflects gender or racial biases, it may continue to favor certain demographics over others, undermining diversity and inclusion efforts. This not only perpetuates systemic inequities but can also expose the enterprise to significant reputational and legal risks.?This one isn’t even a myth.? Stories already abound about issues related to AI-powered recruitment.

Next, there’s the challenge of #data quality. Generative AI relies on vast amounts of data to learn and make decisions. However, if the data is incomplete, outdated, or incorrect, the AI's outputs will be flawed. For instance, a marketing AI drawing on inaccurate customer data might launch campaigns that miss the mark, wasting resources and potentially alienating, and quite possibly offending, customers.

In the realm of creative work, plagiarism and #copyright violations are growing concerns. Generative AI can produce content that closely mimics existing works, raising ethical and legal questions about originality and ownership. Scarlett Johansson and OpenAI ring a bell?? Companies must navigate these waters carefully to avoid infringing on intellectual property rights, which can lead to costly litigation and damage to the brand's reputation.

To mitigate these risks, enterprises must invest in robust data governance frameworks, continuous model validation, and diverse training datasets. Regular audits and transparency in AI decision-making processes are essential to ensure that the AI systems act ethically and accurately.

Myth #2: Generative AI will immediately improve efficiency

While the allure of instant efficiency gains from Generative AI is strong, the reality is far more complex. Generative AI implementation requires substantial time, resources, and expertise, and the path to realizing its benefits is a minefield fraught with challenges.

Implementation Time and Complexity: Implementing Generative AI is not a one-size-fits-all solution. It involves significant integration efforts with existing systems, necessitating a comprehensive understanding of both the AI technology and the enterprise's unique operational landscape. For instance, deploying a sophisticated AI model in customer service requires seamless integration with CRM systems, real-time data feeds, and training staff to handle AI-driven workflows.?Sure, that’s pretty obvious, but we have plenty of data that suggests A LOT of companies are getting this wrong… and we’re in the first or second inning of AI.

Skill Set Requirements: The successful implementation of Generative AI demands a specialized skill set. Data scientists, AI engineers, and domain experts must collaborate to develop, train, and fine-tune AI models. However, the shortage of skilled AI professionals can pose a significant barrier. Companies need to invest heavily in training or hiring new talent, which can be time-consuming and costly.?The Intelligent Enterprise believe so strongly in this that we’ve launched an event in December 2024 called #TechTalentCon.?Check it out!

Initial Investment and Ongoing Maintenance: The upfront investment in AI technology, infrastructure, and talent is substantial. Beyond the initial costs, ongoing maintenance, model retraining, and updates are essential to ensure that the AI continues to deliver value. Not to mention the power costs associated with running these models.?

Data Challenges: High-quality data is the lifeblood of effective Generative AI systems. Ensuring data accuracy, completeness, and relevance requires meticulous data management practices. Enterprises, large and small, need to retool their data governance frameworks to support AI initiatives. Data silos, inconsistent data formats, and privacy concerns can further complicate the process, delaying the realization of efficiency gains.

Change Management: Introducing Generative AI into the workplace is also starting to disrupt established workflows and roles. Employees may resist changes due to fears of job displacement or the perceived complexity of new AI tools. Effective change management strategies, including clear communication, training programs, and involving employees in the transition process, are crucial to mitigate resistance and foster acceptance.

In conclusion, while Generative AI holds immense potential for improving efficiency, the journey to achieving these gains is, as we wrote, a minefield littered with buried explosive devices.? One wrong step and you have big problems.? Enterprises must be prepared to invest the necessary time, resources, and effort into implementation, skill development, data management, and change management. Only by addressing these foundational elements can businesses unlock the true efficiency potential of Generative AI.

2. What Can You Fathom about Generative AI That Others Find Unfathomable?

Like we wrote, generative AI is in the early innings but winning the game will demand that enterprises leverage the technology in ways that others might not even consider. This type of strategizing and planning requires a blend of creativity, foresight, and a willingness to break the mold.

Opportunity #1:

Hyper-personalized customer experiences. Imagine a world where customer interactions are so personalized that every touchpoint feels tailor-made. As John Lennon penned so many years ago, “it isn’t hard to do”.? Generative AI is capable of analyzing vast amounts of customer data to create unique experiences that drive loyalty and satisfaction. While many see AI as merely a tool for efficiency, the real game-changer lies in its ability to create deeply personalized interactions at scale.? Good work is already happening in this area but companies have barely scratched the surface.?

Consider the example of a luxury retail brand using Generative AI to enhance the customer shopping experience. By analyzing purchase history, browsing patterns, and even social media activity, AI can generate personalized product recommendations, style tips, and exclusive offers tailored to individual preferences. This level of personalization can significantly enhance customer loyalty and ultimately increase sales.

In the hospitality industry, hotels and resorts can leverage Generative AI to provide bespoke experiences for guests. By analyzing guest preferences and behaviors, AI can suggest tailored itineraries, dining options, and activities. For example, a returning guest might be greeted with their favorite welcome drink, a personalized room setup, and recommendations for new experiences based on their past visits. This level of attention to detail can turn a one-time visitor into a lifelong customer.? We just had this experience at a Caesar’s Entertainment property in Las Vegas.?Very cool and creepy at the same time….but it’s the future.

Opportunity #2:

Predictive innovation. Generative AI can do more than streamline current processes; it can predict future trends and behaviors, allowing companies to innovate proactively. By analyzing patterns and anomalies in big data, AI can provide insights that spur the development of new products and services before the market even realizes the need.?Consider the possibilities:

• In the automotive industry, for instance, manufacturers can use Generative AI to predict trends in consumer preferences and technological advancements. By analyzing data from social media, industry reports, and market trends, AI can identify emerging preferences for features like autonomous driving, electric powertrains, and connected car technologies. This foresight enables manufacturers to develop and launch new models that align with future market demands, giving them a competitive edge;
• In the healthcare sector, Generative AI can revolutionize drug discovery and development. By analyzing vast datasets of medical research, patient records, and clinical trial results, AI can identify potential new drug candidates and predict their efficacy and safety profiles. This can significantly accelerate the drug development process, bringing life-saving treatments to market faster and more cost-effectively. Additionally, AI can predict disease outbreaks and trends, allowing healthcare providers and public health officials to take proactive measures to mitigate risks.?A few months ago we interviewed a senior executive from United Therapeutics Corporation that was using AI in an effort to identify drugs that could potentially treat – and hopefully cure – rare diseases; and
• In the financial industry, Generative AI can enhance risk management and fraud detection. By analyzing transaction data, market trends, and economic indicators, AI can predict potential financial risks and fraudulent activities. For instance, AI can identify unusual patterns in trading activities that may indicate insider trading or market manipulation. This allows financial institutions to take proactive measures to mitigate risks and protect their clients' assets.

The true, long-term potential of Generative AI lies not just in its ability to improve efficiency, but in its capacity to drive innovation and create unique, personalized experiences. By thinking creatively and leveraging AI in ways that others might not even consider, enterprises can stay ahead of the curve and unlock new opportunities for growth and success.?Which company doesn’t want to be the next Nvidia!?

3. What the Heck Is My Brain Doing to Mislead Me Now?

Cognitive biases are powerful forces that can skew our perception and decision-making. When it comes to Generative AI, these biases can lead to overconfidence, misjudgment, and missed opportunities.

Bias #1: The hype bias.

The allure of being on the cutting edge can blind enterprises to the real, practical limitations of Generative AI. It’s easy to get caught up in the excitement and overlook the hard work, not to mention the cost, required to integrate AI meaningfully into business operations.

The hype bias can lead to overinvestment in AI projects with unrealistic expectations. Enterprises might allocate excessive resources to AI initiatives without fully understanding the necessary groundwork, such as data preparation, infrastructure upgrades, and ongoing model training. This can result in wasted resources and disillusionment when the anticipated immediate returns fail to materialize.

For instance, a company might invest heavily in AI-driven customer service chatbots, expecting an instant reduction in customer support costs, only to find that the bots require extensive fine-tuning and supervision to handle complex queries effectively.?Look no further than the content recently shared at the industry benchmarking event called Customer Contact Week .?Chatbots are on the rise but far from perfect.?If only 50% of the TikTok drive through videos are real, there is a ton of work still left to do.

Moreover, the hype bias can cause neglect of other critical business areas. While focusing on the latest AI technologies, enterprises might overlook essential aspects like employee training, process optimization, and customer relationship management. This imbalance can lead to operational inefficiencies and missed opportunities for holistic improvement.

For example, an organization might prioritize AI-driven marketing campaigns while neglecting the importance of maintaining high-quality customer support, resulting in a loss of customer trust and satisfaction.? Some might refer to this as the #InnovatorsDilemma.

Bias #2: The fear of replacement bias.

There is a pervasive fear that AI will replace human jobs, leading to resistance and skepticism. However, the true potential of Generative AI lies in augmenting human capabilities, not replacing them. By focusing on collaboration between AI and human intelligence, enterprises can unlock new levels of productivity and innovation.? Our research continuously highlights how important the human-in-the-loop concept is to successful deployments.

The fear of replacement bias can manifest as employee resistance to AI adoption. Workers may fear that AI will render their skills obsolete and lead to job losses. This resistance can slow down AI implementation and hinder its integration into existing workflows. To address this, enterprises must emphasize the role of AI as a tool to enhance human performance rather than replace it.

Additionally, the fear of replacement bias can lead to underutilization of AI capabilities. Companies may hesitate to fully deploy AI technologies due to concerns about potential job displacement and backlash from employees and stakeholders. This cautious approach can prevent organizations from realizing the full benefits of AI. To overcome this, enterprises should invest in retraining and upskilling their workforce to work alongside AI.

In conclusion, cognitive biases such as the hype bias and fear of replacement bias can significantly impact the successful integration of Generative AI in enterprises. By recognizing and addressing these biases, businesses can adopt a more balanced and strategic approach to AI implementation. This involves setting realistic expectations, investing in comprehensive training and change management, and fostering a culture of collaboration between AI and human intelligence.

Conclusion

Generative AI holds immense potential for transforming enterprises, but realizing this potential requires a critical, informed approach. By asking the right questions—What beliefs are false? What can we fathom that others can't? How are our biases misleading us?—we can navigate the hype and harness the true power of Generative AI.

Stay tuned for more insights and join us at the GenAI & ML Tech Stack World Series event this October in Los Angeles. Together, we’ll explore the future of AI in the enterprise and discover how to turn today's questions into tomorrow's innovations. ??

Ken Fisher Melissa Lattman Caroline Jacklin Mercedes Mayfield Rob Shannon Customer Contact Week Customer Management Practice