Zohar Bronfman is the CEO and cofounder of Pecan.ai, a predictive analytics platform built to solve business problems.
getty
A prediction for 2024: The majority of companies’ generative artificial intelligence (GenAI) initiatives will fail.
It’s a sobering thought. The GenAI buzz has been tremendous, and the investment has been massive. Champions of GenAI projects have expended a lot of personal capital on getting buy-in, and leaders are waiting to see meaningful business impact.
But those who know the history of data science and AI implementation can see the speed bumps coming on the highway of hype. And yet, despite those statements, my outlook for the future of GenAI is genuinely positive.
Let me explain why 2024 will challenge early adopters of GenAI—and why it’s not too late for them to refocus their AI efforts to realize far greater business value.
Hard Truths About Generative AI’s Business Potential
It’s been phenomenal to witness the rise of new generative AI tools. Like everyone else, I’ve been awed by what these powerful models can create.
At the same time, I’ve shared concerns about generative models’ accuracy, the potential for bias and ethical and regulatory considerations. Those are critical issues. But another major concern is that the buzz around GenAI has distracted many from deep scrutiny of these technologies’ true business potential.
McKinsey projects that GenAI has the potential to contribute as much as $4.4 trillion to the global economy each year. While that’s a staggering number, it’s still just a quarter of the $17.7 trillion that advanced analytics, traditional machine learning and deep learning can contribute.
In reality, the more familiar ML-based methods we already know and love still represent massive unrealized potential—if companies fully embrace these well-established technologies. In North America, 42% of companies have not adopted AI and ML.
Of course, humans tend to be distracted by new, shiny objects. So, we’ve seen many companies jump into GenAI initiatives, even as they neglect essential machine-learning use cases. Those overlooked ML projects could have a far more rapid and significant impact on their business outcomes.
Data science history is repeating itself.
Here’s where that knowledge of history comes into play. When data professionals and businesspeople become too fixated on AI models themselves, failure is too often the result. We’ve seen this tendency before. It’s why nearly 90% of data science projects have failed in the past.
The approach of many GenAI adopters right now is like opening a toolbox, picking up a shiny wrench that looks fun to use, and then wandering around your home to find loose bolts to tighten. But are they even essential bolts? Is this the best use of time and money?
A much more efficient and rewarding approach would be to determine which repairs are needed and then pick the right tools for each job. Focusing on the latest, shiniest new algorithms leads businesses astray. We’ve seen it happen before. Instead, companies should rigorously pinpoint critical business challenges and then identify AI tools that best address those issues.
Generative AI success will be more challenging in the short term.
Other factors will affect generative AI’s success in the year to come. As the hype cools, reality will set in.
The ongoing GPU shortage, the cost of compute for GenAI models and the environmental impact of GenAI usage will all take a toll. Companies may also experience the consequences of bad actors’ harmful use of GenAI, incurring additional costs.
With all of these potential issues, many early adopters of GenAI are likely to find that their initial experiments aren’t providing the ROI they hoped to see. If the results are disappointing—as I predict they will be for many—it’ll be time to explore new strategies for implementing GenAI into enterprises.
Twinning Is Winning: Generative AI Plus Predictive AI
I’ve called GenAI and machine learning (or, more broadly, predictive AI) “twins.” They’re fraternal twins, to be sure, but they are complementary. I believe companies that find ways to use these twin technologies together will see the greatest impact in 2024.
I’ve written here before about how marketing, in particular, can achieve new efficiencies and create a better brand experience through the use of generative and predictive AI together. For example, an e-commerce subscription provider may recognize rising customer churn. It’s not a real solution simply to use GenAI to create lots more email messages or web content. Instead, the generative and predictive AI twins offer a far more sophisticated solution if combined. What could that look like?
With machine learning methods, the company can predict which customers may be at the highest risk of churning. GenAI can’t do that because it can’t work well with business data that’s primarily numbers in a tabular format. But with customers segmented according to their likelihood of churn, the company can decide who should get which discount or offer, allocating their resources efficiently (and not spending money on happy customers predicted to renew).
But machine learning can’t do it all. GenAI comes into play when it’s time for the company to reach out to at-risk customers individually. GenAI helps make it easier and more efficient to craft personalized messages and content to resonate with each customer. Through precisely targeted and personalized messages, the maximum value of both forms of AI can be reaped, ideally resulting in an efficient, effective customer retention effort and high ROI.
This is where I believe we should turn in 2024, as GenAI’s shine fades and we recenter ourselves around the real “why” of AI in business. Adopting AI isn’t just an excuse to experiment with the latest new tool. Instead, AI should enter the picture because you’ve identified major challenges and opportunities for your business that can be addressed with its help.
Then, with those key tasks in mind, you can select the best tools for the job. That will likely mean using multiple tools in coordination, uniting exciting innovations and well-established technologies to achieve even greater success.
Keeping this true "why" of AI at the forefront will help refocus AI initiatives on making a meaningful business impact in 2024.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
