Companies, wake up! The biggest change AI brings to companies is not technological but conceptual.
Machine translation from the original post. Not reviewed by a human translator. Historical claims may no longer be current.
Companies, wake up! The biggest change AI brings to companies is not technological but conceptual.
You can give employees the best model, connect it to all the right tools, and still see them work almost exactly as before.
Our ability to reimagine how work should look when this technology exists is an unnatural and non-trivial act, and requires ongoing engagement and changes of habit. This can't be solved just by giving access to tools and expecting that "everyone will figure it out."
That's why the term FDE became popular in this AI world. The idea is not only to develop a product and hand it to the customer, but to put strong technical people deep inside their work environment, understand the real problem, and build with them a solution that works in reality.
In my opinion the same principle should exist inside companies too.
Every organization that wants to implement AI seriously needs people who function like internal FDEs. They need to enter the teams, sit with the people who do the work in practice, and do a Drill Down to the level of day-to-day tasks.
Not just ask which AI tool you're missing, and not just run some superficial workshop introducing this or that tool, but dive deep into every team and understand why you do this task this way at all, what can be automated, what can be changed, and which process no longer makes sense in a world where AI exists.
This also requires AI Champions from within the teams themselves, regular sessions and trainings based on each team's real work, and not another general training on how to write a good prompt.
Because if right now the goal of these people in your company is to build AI for the employees - it will most likely fail. It's building tools that nobody will use.
The goal is to change the way employees think about their work, and that requires much more than developing tools.
This is probably one of the reasons that implementing AI is harder than buying software. New software requires the organization to learn a new tool. AI requires it to relearn what humans even need to do and mainly what not.
On the bright side, this is a problem all organizations are dealing with right now, if that comforts you :)