Is the user experience as important – when all the marketing out there seems to all be the data models underpinning the AI?

In the enterprise landscape, it’s not just the LLM models that matter. In fact, I’d argue the underpinning LLM models matter less than two other factors: (1) the product design and (2) the data that powers the AI. 

Business applications are more complex than consumer ones because a number of systems, processes, and people come together to achieve a certain outcome. Most businesses ensure efficiency through standardized workflows and repeatable processes for employees to follow. And so it’s within this context that AI needs to exist. As a result, it takes thoughtful product design to properly embed AI into a business's internal ecosystem.

The actual data - both the quality and quantity - is the other factor to consider.  As the saying goes: garbage in, garbage out. When I led Google Ads product marketing, I saw how a lack of data (both breadth and depth) could limit an advertiser's ROI when using Google AI campaigns like Performance Max. The best advertisers fueled the AI with lots of high quality data across targeting (first party signals), creative (uploaded tons of assets from text to video), bidding (shared customer lifetime values for value-based bidding). Now that I’m in the Revenue AI space, I see the same thing — sales teams are fundamentally stunted when they fuel AI with CRM data that’s collected from manual seller inputs, instead of automatically capturing all of the customer interaction data using AI directly.