With the full swing of FY24 planning underway, enterprise leadership teams are taking a good, hard look in the mirror to assess where they stand, where they want to go, and how they’ll get there. Unsurprisingly, much of this corporate introspection is focused on generative AI.

In the conversations I, and likely you, have had on the topic, one thing is abundantly clear: generative AI ambitions for the future are ubiquitous and grandiose. There is no shortage of excitement. But there is a shortage of thoughtful planning and execution.

Don’t get me wrong, I’ve seen countless companies experiment and test generative AI applications. C-suite executives ponder the use-cases where generative AI applies. Developers have begun familiarizing themselves with the LLM APIs. HR teams have begun posting openings for AI-specific roles.

So who will lead these initiatives across organizations? Who is the authoritative voice evaluating generative AI investment and adoption? What about navigating the newly arrived and imminent legislation in the US and the EU, and ensuring compliance with these laws?

Currently, there is a void in dedicated AI leadership, management, and technical talent across industries. That needs to change before we can see the promise of generative AI realized. It’s not enough to want to become a leader in AI. You actually need a leader to help you get there. Dedicated AI leaders can prevent a “part-time” AI strategy, one that takes a back seat to other priorities within your organization and limits hands-on experience with AI solutions.

At data.world, we have Brandon Gadoci, our VP of AI Operations. Brandon has been with us since the beginning of our company and has a very entrepreneurial spirit. We realized that we needed someone highly creative in this role, and Brandon says this is his favorite role to date. He leads the creation of bespoke AI-powered productivity tools. We have nine tools now, eight of them internal (Prospect Researcher, Account Researcher, Landing Page Copy Generator, RFI Writer, Strategy Primer, Email Campaign Writer, Docs Explorer, and Product Brief Maker) and one of them external (but used a lot internally too): Interactions with Archie (you can check it out here). We’ve already seen an incredible 25% increase in productivity among our employees that use these new tools — and it’s truly just the beginning.

Brandon has achieved this productivity lift by systematically going from department to department looking for the highest ROI bottlenecks to develop these tools. He’s been writing his own series on this at bgadoci.com and on his many LinkedIn posts, which I encourage you to tune into. You need someone like Brandon in your organization to do the same. I remember when I was earning my MBA at The Wharton School, I was first turned onto the book The Goal by Eliyahu M. Goldratt about finding your “Herbies” or bottlenecks. Thanks to investor Ralph Mack, we implemented The Goal’s process at my prior company, Bazaarvoice, to have an incredible outcome in our customer implementation process, resulting in much faster throughput, better gross margins, more upsells, and a higher customer satisfaction rate. That is essentially what Brandon is doing at data.world with these newfound productivity superpowers, and it is creating employee joy throughout our company. As my good friend, fellow entrepreneur, and incredible AI author and podcaster, Byron Reese, once said to me, “It is actually inhumane to make someone do work with inferior technology, struggling in the drudgery of slowness.”

In 2023, we went all in on AI for our customers and built a dedicated AI Lab to explore how AI can augment our data catalog platform, and the impact it can deliver for enterprises. Juan Sequeda is our Principal Scientist at the helm of the lab. His vast personal talent and expertise — coupled with the freedom to experiment and explore AI integrations — allows impactful assessments that support our customers’ unique AI strategy and execution.

In one of the most significant endeavors for our AI Lab, Juan, alongside Bryon Jacob (our CTO and Co-Founder) and Dean Allemang (our Principal Solutions Architect), published a benchmark (full paper linked here) on LLM accuracy measuring improvements delivered by a knowledge graph architecture. The study uncovered a crucial finding: LLMs struggle in answering business intelligence questions of varying complexity in SQL databases. Knowledge graphs, however, delivered a massive 300% increase in accuracy, backing the main hypothesis that when data is organized and connected through a knowledge graph architecture, AI systems have a better organizational “fact sheet” that can improve their accuracy and earn trust.

But the work didn’t stop there. dbt Labs took up the mantle and went on to verify and replicate the findings, publishing their report Semantic Layer as the Data Interface for LLMs. The initiative, led by Joel Labes and Jason Ganz, came to the same conclusion: layering a Semantic Layer over the top of your data allows enterprise teams to ask diverse natural language questions and get responses that you can trust. [Note that data.world partner ThoughtSpot is hosting Juan, Joel Labes, and Mode Analytics founder Benn Stancil on January 31 at 10am PT for a discussion on this that you can register for here.]

The collaborative work for the future of AI cannot be overstated. It starts with investments and commitments within our organizations. The work requires a leader that asks the right questions, implores their teams to find the answers, and then leads them to build stronger products and solutions.

As enterprise AI initiatives mature, a new, dedicated cohort of AI executives — whether it be a Chief AI Officer, Head of AI, or VP of AI — will become a strategic imperative for enterprises. Site Reliability Engineers and Database Administrators used to be novel; now they are standard parts of a working engineering team. Formal AI roles will follow suit.

These AI champions will be the lynchpin to AI experimentation, development, and implementation. From evaluating and operationalizing AI for industry-specific use cases, to managing the security and compliance of AI applications, they will be critical to rolling out enterprise AI strategies.

2024 will be the year of the AI executive. My advice to other CEOs: write the job description for an AI executive role. The productivity lift you’ll get will enable you to stay ahead of the competition, improve your margins, and increase the joy of your employees as they get more accomplished in exciting new ways. It will be a very exciting journey for your company!