Generative AI: Elephants in the Room

The year 2024 is poised to be a landmark for transformation as organizations refine their generative AI strategies, explore potential use cases, evaluate third-party vendor capabilities, and integrate these elements into their business operating models. The challenge is not only in crafting a coherent strategy but also in its execution.

In our industry, the initial inclination has been to prioritize technology. Early stages of generative AI centered around conceptual proofs and modest projects. This excludes giants like FAANG, whose resources dwarf the typical investment. The rush to leverage language models, vector databases, and prompt engineering saw many trying to kickstart their journey into generative AI without addressing future scale and integration challenges.

Optimization Formula: Variability

An optimization formula can be defined as a function and the variables that determine business outcomes. While we've defined these in the past, with technology waves such as big data. The equation included variability and for the most part it was understood and grew slowly.

With generative AI, the impact of the variability variable is much larger. The speed and size of change being created by generative AI has significantly more impact. It may appear small when the focus is on a single use case or project. When there is an attempt to scale a project, or scale generative AI across an organization the dimensions of generative AI variability has larger consequences.

Elephants in the Room

The metaphorical 'elephants in the room' while substantial are often not understood, minimized or not prioritized. There are consistent patterns in repeating the same mistakes from previous paradigm shifts. Also, let's not kid ourselves, data governance, data quality, change management, process management, and solving the human element, etc. are extremely difficult and not getting easier. The success for these projects are ~20% or less.

The 'elephants' become much harder to overlook when moving towards scale and speed or operationalizing a platform. These elephants could also include legacy mindsets, technical debt, siloed environments, or a focus on technology over business outcomes. What are your elephants?

Conclusion

Successful deployment of generative AI requires CDOs to cultivate teams capable of innovative thinking, imbued with a growth mindset, and meticulous about the nuances of building and scaling a generative AI platform. We must learn from past errors. I repeat, a recurring theme in generative AI initiatives is a fixation on the technology itself, rather than its business application or impact.

We are in a rapid growth phase and all that comes with it. Everyone needs to get faster at test, and if we fail, LEARN, then repeat. CIODive has an excellent article on this topic. Mature your processes to continually evolve to build the skill and expertise in operationalizing generative AI platforms for scale.

Generative AI harbors the potential for revolutionary change. I eagerly look forward to continuing to explore and sharing this journey with my peers, as we unlock the myriad possibilities it presents.