Enterprise AI projects are stalling at the prototype stage and the hype around #generativeAI is subsiding. Despite the myriad of tools, promises, and impressive demos, few companies are making any real progress.
Enterprise AI feels like smoke and mirrors. A lot of talk, not a lot of real progress.
This is the observation I shared with a US tech investor who also works in the enterprise AI space, and he agreed wholeheartedly.
We have both found there are two enormous barriers: the legal department and the IT department.
The legal department wants to lock everything down and eliminate risk. Just like the introduction of the automobile, they want a person with a red flag to walk in front of the vehicle to make sure everything is totally safe, thereby eliminating the core benefit of the technology.
The IT department wants to run everything in-house but lacks the technical or commercial skills to scale. For decades, CIOs have follow the Microsoft blueprint and built their organizations around the windows/sql/intel stack.
In doing so, they actively suppressed cloud, Linux, and Python, which are now the three fundamental building blocks of an enterprise AI solution.
Learning new skills takes time, but every IT department I know has a backlog of projects a mile long; they simply don’t have enough capacity to learn and integrate a new set of technologies.
So we can’t be surprised when busy, overworked IT departments get a new toy and don’t know what to make of it. Anything they do create goes through extensive legal review.
This is just the way that organisations actively protect themselves from change. It’s a self-defence mechanism to slow things down, which is why enterprise AI is not living up to expectations.
Innovation in Enterprise AI won’t come from enterprise.
It just can’t.
