Perspectives: Ontology Is the Hidden Force Powering Business Success with AI
- Share via
You may have caught Palantir’s Q4 earnings call last week. Maybe you didn’t. Either way, I found something fundamentally thrilling about what Alex Karp and his team had to say. And no, I don’t own shares of Palantir stock, which has skyrocketed with remarkable growth across all the right metrics.
One word they said in particular piqued my interest. A term rarely heard outside of philosophical discussions on existence and reality, it refers to the structure of being and the relationships among its features. That word? Ontology. And I anticipate it will soon become a buzzword in the AI lexicon.
On the call, the team credited “ontology” as the reason for Palantir’s success: an ontological system concretized in the form of a technological product. This solution connects the vast array of business or individual data in a way that allows artificial intelligence to fully leverage the knowledge within it. To make AI data-centric, an ontology maximizes the value of the data by enhancing the information a model can extract from the data inputs. This foundational work in a prediction system enables the output of downstream AI, or the agents designed to carry out specific tasks, to perform in intentional, explainable, and reconfigurable ways. And as a bonus, this foundation makes it easier for users to iterate and innovate on how AI is used and consequently the overall performance level it reaches.
A few weeks ago at a panel organized by LA Times Studios, I shared that the most surprising development in AI is that the intelligence model layer, or weights (not its training), has fast become an essentially free commodity. At the same time, the quality of these models continues to improve, while the application layers are being refined and perfected. For a business or individual to unlock the full potential of AI, their data – and the knowledge or information contained within – must be organized through an ontology that aligns with the intelligence models required to carry out the business or productivity goals of a prompt given in the common language of its users.
Ontologies relate to defined bodies of knowledge, and while they may overlap, successfully implementing purpose-built AI for specific domains such as production and supply, finance and accounting, or sales and marketing, requires the right productized ontology technology.
As my expertise lies in consumer economics and marketing, I’ll offer a relevant example from that domain. Consider the process of orchestrating an advertising campaign for a brand. Once the initial strategy and creative concept are defined, a traditional creative content production team assembles images, videos and ad copy in various formats for placement in specific media. This process is slow, laborious, and typically results in only a few ad variations. However, an experienced advertising professional can prompt an AI application that writes code and sets up the necessary IT infrastructure, enabling the creation of virtually endless variations of the ad. But that’s not the big idea. The real breakthrough lies in the AI’s ability to categorize the ad’s features, using ontology-based technology to relate those features. For example, it can link detailed data about the ad copy, font, imagery, format, color schemes and more to the target audience, the context in which the ad is viewed, and the behaviors and touchpoints that track an entire population’s journey through the marketplace.
These components, related to one another taxonomically in their stored state, are tracked to measure performance and to reuse elements as inputs for AI even when they are only available in incomplete and more privacy protecting aggregates. Purpose-built AI can fill in gaps and assess the impact of various aspects of consumer experiences. The vast and varied knowledge generated through this process becomes the essential input for producing new creative ideas, strategies, campaigns, and content. Such an application dramatically reduces labor costs, increases speed and eliminates friction and entire steps in prevailing processes – while delivering far more impactful execution.
Ontology will be the key differentiator in the race to harness AI’s capabilities. Early investment in the right ontological framework will determine who leads in the AI ubiquitous world, enabling smarter, more efficient solutions that drive transformative business results.
-Michael Cohen is a technology business chief executive and globally recognized leader in AI, data, and predictive computation. He specializes in transforming businesses with data-driven management systems and is most experienced at bringing consumer data and AI technologies to market.
More Stories
Sedona 2-Acre Getaway Retreat
Legends Peak
Gated Development in Calabasas
