Are you saving time with AI?
Are you thinking about how to best apply AI to save time in your organisation?
Or are you looking to increase efficiency by using AI?
Or are you looking to increase efficiency by using AI?
These two things are very different.
Read on, there are more things you may have missed.
Over the last 24 months, we have all seen and read too many articles about how to save time using AI, GenAI, AI agents, etc. There has also been a huge debate (still raging) about jobs being lost when AI is introduced.
Yes, some companies have cut jobs because of AI, but decision makers in many organisations are not even aware of how and where AI is applied in their organisations, which means they haven’t yet been able to discuss any potential savings or gains, let alone use that as a reason to cut jobs. Yet. If they ever will.
I know of several such examples where AI is used for mundane things as well as bigger, more complex tasks, but where senior management may not be aware of exactly how, where and why.
Apart from the obvious risks created if AI is used without senior management having good enough insight, there is also a risk that benefits are misunderstood or miscalculated, leading to misguided and uninformed discussions and decisions.
Whenever we hear of an organisation that claims to have saved time by using AI, we should stop and think.
Let’s take an example
A software product company produces and sells complex software.
One of their senior engineers has used AI for a long time to write and check code, create tests, etc. Additionally, that engineer has also created a few autonomous AI agents to help with some related tasks.
Did it save time? Potentially. Not necessarily.
Did it save money? Same thing there.
It is only saving them time if something that was already being done is now done more quickly. However, if new things are being done that weren’t done before, that does not constitute saving time.
If that senior engineer does some really great things that were not done before, that does not mean time was saved.
If product quality went up, that does not mean time was saved.
If the need for engineers in the company decreased, but the company still kept on hiring, was time saved? Was money saved?
Sure, it may be a good decision to keep on hiring, but was the decision to contunue hiring at all influenced by the new savings? No? Doesn’t that mean that the company is making uninformed decisions? Ouch.
Sure, it may be a good decision to keep on hiring, but was the decision to contunue hiring at all influenced by the new savings? No? Doesn’t that mean that the company is making uninformed decisions? Ouch.
Did the company mange to realease new features more frequently? Nice! …but what was the quality of those features; did they add a lot to the existing tech debt? Was time saved?
Let’s look at this from another angle
In some ways, a software company can be compared to a network of roads; a traffic system. In a software company, some things always take time because there are just too few teams or too few specialists in a certain area.
The number of teams is limited by budgets.
Translated to a traffic system situation, traffic is limited by the number of lanes on the roads. ’Induced demand’ is a concept often used in traffic planning. It means that people just travel more if there are more lanes. The congestions remain. Time is not necessarily saved.
In economics, there is a similar thing called Jevons’ paradox.
In short, it says that if there are technological improvements that lead to improved resource efficiency for a certain resource, this may just lead to people using more of that resource, not saving it.
What does this mean for
software product development?
In order to steer the organisation in the right direction while avoiding risks (economical, legal, security-, etc), senior management should always have good insight into what is going on in the organisation.
AI has become an area where it is sometimes difficult for senior management to know and understand as much as they should. This is even more difficult for the board, whose job it is to oversee and steer things on a higher level, because they are further away from the details.
So, when someone says that AI is used to save time, we should be a little sceptical or at least ask questions.
Something great and very interesting
may have been done, but
it may not necessarily have saved time.
It might just be that more was done instead. It might be good things, sure, but that doesn’t mean time was ’saved’.
Did it save money? Not necessarily.
On top of this, the use of AI adds relevant costs in itself. There are many examples of organisations that started using AI for as much as possible but had to act quickly when the costs went through the roof.
To all senior managers and boards
Don’t stop asking your usual questions just because someone says ’AI’ or claims to have saved time or money.
Things may be great and there may be a good plan, but you should at least ask.
Take care, everyone!
Björn