We’re all great at wanting to give things a crack, it’s just human nature to look at something and claim “I can do that”.
While that’s admirable for my 10-year old’s cricket aspirations, we need to be smart when we are considering building infotech solutions from scratch or buying something that someone else has created.
This has even more relevance the fast paced emerging AI space.
Our experience has shown that it takes many iterations to develop a piece of AI tech in order to really nut out how to optimise it and get the most out of the solution – it’s not just a case of scribing out some python code and then deploying it.
There is a middle step in there called “trial and error” and unless you’ve got a serious amount of time to tweak, adjust, discard, and build again, then you’re probably better off getting a solution that is already doing the job you need.
This will allow you to fast track your time to getting the outcome you need and free up your tech whizz for other bespoke solutions that must be done internally.
This doesn’t mean that you will have to settle for an ill-fitting out-of-box solution as any AI solution partner worth their salt will be able to work with you to optimise the solution to your specific needs.
Plus, you will get all the learnings they’ve gained from deploying their solution in many more scenarios than you would ever be able to replicate in your single organisation – group learning is hugely advantageous in AI – after all that this the whole point!
So, think hard before you get into DIY AI, as there may be smarter options out there that can shortcut the development cycle and get the outcomes you need faster.