So, the future is here. We now have access to a magic box that can do all sorts of tricks. However, the question we're now being asked increasingly often (and also ask ourselves) is "Now what"? How exactly do I use this thing to help me with my work? As you can probably imagine, there's not a one-size-fits-all answer.
The biggest issue in understanding how AI assistance can help you is that so often it's spoken about in the abstract. Lots of talk about how amazing it is, but often without much detail on exactly how.
There's a couple of reasons for this. Firstly, even the providers themselves aren't necessarily certain of the answer to that. Not that there isn't one. Just because we're still at the start of this journey, so we're still very much in the "work it out as you go" stage. The second is the sorts of things people could use it for are so varied, both in terms of scope and in terms of the time that you personally could save in using it, that it's pretty much impossible to pitch it to everyone at once. Sometimes, it's very simple. The advantages it can bring to the development side of our business in particular are so obvious that it only takes minutes to grasp the potential. But in some lines of work and in some roles, it will be far less clear cut than that. In some cases, it may only be the fringes of the job it can help with. Still better a lot better than nothing, but this seems to be where a big source of frustration lies when it has been sold as a miracle machine.
How to roll out AI assistance across your company
With that in mind, how does one go about rolling AI out to an organisation when it isn't even necessarily clear yet how exactly you want everyone to use it? Our advice here would be to do this slowly rather than trying to force it on everyone at once. In our experience, some people are very keen and proactive in experimenting with the tools available. Others want to be told exactly how they're expected to use it, and won't be sold on the idea until it's demonstrated to them in real terms for an application that is specific to the way they work personally.
Therefore, we would suggest concentrating on that first group to start with. Encourage them to proactively try things out. If they see some new app that could help in their particular case, then let them give it a go. Once they've got a grasp on what's available and how that works for their department, then use that person to explain it to the others in terms that can be directly related to the existing processes.
Do you want AI with that?
In many ways, the AI companies are not helping themselves here. Because they see a gold-rush coming, they're throwing everything at the wall to see what sticks. That makes sense from a business perspective - if you come up with that killer application, you're going to be making a lot of money. The problem is, a lot of that stuff isn't sticking. This is particularly true outside of the business-context. Everyone is getting AI flung at them from every direction, and a lot of it is just a token effort to jump on the bandwagon. It therefore isn't surprising that a lot of people are just bouncing off, or even becoming actively hostile towards it. In this context, it's entirely understandable if their reaction to new "helpers" popping up all over the apps they use without clear instruction on how they're expected to use it is "Go away and let me get on with my job the way I've always done it!"
The other consideration here isn't just whether AI can speed you up in a certain area, but whether you should be letting it. Take emails, for example. This is, in theory, something pretty much everyone could use AI to assist with. It can churn out whole essays in minutes if you want it to. The flip side of that is AI generated emails can be very generic. It is literally what AI does - provide the most predictable set of words. "Given this sentence so far, what is the most likely word to follow the word I've just added". With a bit of hand holding, that can help someone that doesn't find writing emails easy to produce something a lot more professional-sounding. However, what it can also do is completely remove any personality from your emails. Is that the perception of your business that you want to convey? In some cases the answer will be yes, and in some cases no. In a world where emails are becoming more uniform, is it better to stand out from the crowd? If every company is using AI to produce everything, then the main points of difference between you and others providing the same service will be price, and how you present that work.
As ever, there's not simple answers to be had here. However, it's definitely worth thinking about these things and then coming up with company policy for this stuff, and be prepared to evolve that policy rapidly as tools change and you get feedback from staff on what works and what doesn't. Work with the people that are keen, and listen to the people that are resistant. They are probably both right, at least to some degree. Then sack all the resistant ones and replace them with a Claude subscription for just $20 a day. Hang on, they've just changed that to $40. Wait, $60 now. Hmm... can someone remind me how to write an email myself again? What do you mean we sacked everyone because Claude said it could do it all - get them back!