Not really. Though pushed to chase “Generative AI” thanks to the hype around it, sadly, most tend to focus solely on the shiny end-products without understanding the hard work, innovation, and data engineering that precede such success.

The generative AI landscape is peppered with eye-catching headlines of colossal acquisitions by major companies. For instance, the news on DataBricks’ $1.3 billion purchase of MosaicML, Accenture acquiring Flutura, and Snowflake’s acquisition of Myst AI, did the rounds of the startup world recently. While these transactions signal the importance and value of generative AI, do the CEOs pouring their all into it really understand the infrastructure and tooling that goes behind it?

“They want generative AI. To them, Microsoft is the generative AI Company, and ChatGPT is generative AI,” remarked AI advisor Vin Vashishta, in a LinkedIn post. 

Business leaders view technology giants like Microsoft as the epitome of generative AI capabilities. Every company wants to be Microsoft and wants a ChatGPT under its belt. However, this simplified perception disregards the critical role of underlying data engineering infrastructure and tooling.

The mirage of a generative AI product

A disconnect exists between the CEOs’ visions for implementing generative AI into their companies and the technical intricacies that underpin its successful deployment. There are various challenges that come with building a generative AI framework, from deciding if it’s even plausible to be used in the company to deciding how to deploy it. Interestingly, CXOs find this hard to understand. 

CEOs pushing their teams into generative AI without comprehending the underlying infrastructure and tooling requirements is because of the push they get from the investors and market trends. Even then, it highlights the need for education and collaboration between business leaders and data professionals to ensure successful integration.

CEOs and top-level executives are often captivated by the allure of generative AI amid all the funds and money that this technology is minting for companies around them. Skipping literacy about the technology, the infrastructure required, the expertise required for the team, the availability and ethical boundaries around using the data is often ignored. The push is directly given to the team to build a so-called “generative AI product”.

It is the data teams that are tasked with the Herculean implementation of generative AI, grapple with complex mesh structures, data governance, legal considerations, and contractual obligations. While these technicalities are crucial for successful AI integration, they are often lost on CEOs, who are primarily interested in the final output rather than the steps to achieve it.

Who is to blame?

This discrepancy between the team often stems from the communication gap between business leaders and data professionals. CEOs use a different language, emphasising results, market positioning, and profits, something that the data teams do not really care about if the process of building models is not streamlined. This disconnect leads to misaligned expectations and potentially frustrating interactions.

Vashishta gives an example in his post that when data teams return from conferences brimming with enthusiasm about cutting-edge infrastructure and tooling, business leaders fail to share their excitement. “To CEOs, these concepts appear as incomprehensible dots on a Gartner report, rendering them unenthusiastic about investing in the necessary foundation.”

On the other hand, the success of technical and data teams hinges on turning data into actionable insights that help leaders drive the business forward. To empower leaders, it’s crucial to provide accessible information instead of using technical jargon. Generative AI offers valuable outcomes, but failure to deliver insights can impact funding. In the end, the business leaders care about the ROI. 

AI literacy for CEOs

When data literacy was the talk of the town, CEOs were saying, “we are a data literate company”. And now, the CEOs say that they are “AI literate”, when arguably they may not be. To bridge this gap, it is imperative to educate business leaders about the core aspects of generative AI implementation. CEOs must comprehend that investing in AI engineering infrastructure and tooling is the bedrock on which successful generative AI initiatives rest. Being data-literate is one thing, but it’s time they became AI literate too.

By providing business leaders with a clear understanding of the steps involved in AI development, including data collection, data preprocessing, model training, and validation, they can make informed decisions and set realistic expectations with their teams. 

For the AI teams, it is important to understand that infrastructure isn’t actually a C-Level task. Generative AI is a strategic goal, driven by C-level, but AI infrastructure and quality are tactical and should be handled by the middle management. C-Suite trusts middle management to make the right decisions and convey the importance of data quality and infrastructure. Generative AI is a means to an end, not a strategic goal itself for the CEOs. That is expected to remain the case till the time they get AI literate.