Projecting the Adoption of LLMs
I’m going to go a bit contrarian to other vendors for my 2023 forecast. In that, despite the massive promise of LLMs and AI, meaningful adoption is still...
I’m going to go a bit contrarian to other vendors for my 2023 forecast. In that, despite the massive promise of LLMs and AI, meaningful adoption is still 18+ months out.
Because:
Adoption is an adaptive process in which organizations need to rationalize their internal value systems around what AI means to them and what use cases they’re willing to embrace.
The binding constraint to this adaptation, and thus adoption, is the lack of unified and transparent data across all the applications that touch CX. The data needs to be readable by both machines and human stakeholders in the organization in the context of their role and scope of responsibility.
In other words, machine learning models require a lot of data to be trained accurately, and humans need to make sure those models stay in check and fully comprehend the impact on the organization (and not just the AI itself).
Yet, vendors continue hawking closed box "AI" systems and point BI tools.