Yesterday, I had a great conversation with a friend about how the AI assistants are making us more efficient in our day-to-day job.
At some point in the conversation, he said:
Now you can do the same job with fewer people, why would companies hire more people?
I think this argument has been repeated though the IT industry for a long time, and I think it is wrong.
It is true that they are making people, teams, and organizations more efficient. But this happened before, and we didn’t see a reduce on global spend, but an increase.
The Cloud
I started using AWS in 2014, back then there was the claim that with the Cloud you didn’t need huge Sysadmin teams, and it could reduce your IT spend by a lot.
In fact, it kind of happened… more or less.
Let’s assume that the Cloud was a that huge competitive advantage, and we have two companies.
Company A. Continuous working as always.
Company B. Adopts the Cloud and DevOps practices.
Based on the claim, B will become more efficient, will spend less, and will be able to offer a better product, at a reduced cost compared to company A.
After sometime, some stuff can happen:
Company A adopts the Cloud and gains the same competitive advantage as company B.
Company A dies, leaving in the market all the companies with the same competitive advantage.
So, as time passes, the competitive advantage is no longer a competitive advantage because of all the players in the industry either adopted them or died.
The same will happen to AI assistants
Right now, we are feeling super productive.
We can debug systems faster by asking some LLM to explain to us what’s happening based on the logs. We can ask them to scaffold a code to do a form. Not only that, but we can ask them to help us migrate a code base from one platform/framework/language to another.
So, adopting AI assistants provide us a competitive advantage over the ones that didn’t adopt it.
Until everyone adopts it.
And then what?
You will re-invest on people, you will grow the teams, and your competitive advantage might be the human capital again (or it always has been?). Their know-how and how fast they can learn will become the competitive advantage again.
Efficiency isn’t a competitive advantage when everyone is the same efficient doing the same thing.
What I don’t fully know is if the investment will be back to developer teams, or it will be another company function.
Domain complexity
I see how Copilot helps me do some CRUDs, forms, lists, … in a crazy fast peace. If the cost of doing those basic operations drops a lot, how can we protect our products if they are that easy to copy?
I believe we will be creating products with a bigger domain complexity than before, and the go-to-market, marketing, branding, and pricing models will gain more weight.
Product alone won’t (isn’t?) a competitive advantage, we need the full set of functions to work together to create a compelling experience and solve real user problems.
Siloed teams, like developers and marketing teams not talking/collaboration (for example), will become a bigger problem.
That domain complexity will require several functions and roles to collaborate and exchange knowledge more than before. The data and know-how of the customer will become more relevant to bet on more complex and riskier initiatives.
The low-hanging fruit talks are gone. Yet, the expectations on the baseline of features will keep increasing.
Huge product prioritization will be required as developing low value expected features will be easier, and the effort of working on uncertain-high risk task will increase.
Otherwise, we will fall into the trap of developing fast, shipping fast new features, while not providing any value because everyone else have the same set of features.
Distinguish products, and understand their value to make a decision on what to use will become crazy hard. People that know their customers and know how to communicate and capture value will become more important.
AI assistants and big organizations
At first, I thought that smaller teams could go faster, because of being more efficient and able to produce more with fewer people, and that could be a huge competitive advantage compared to bigger organizations.
But I think AI assistants enables different aspects for small and big organizations, depending on their needs, bottlenecks, and flow efficiency problems.
My hypothesis is that AI assistants will (or are already) enable big organizations to reduce their team to team dependencies, one of the main causes of huge team cognitive load, and it could help them to innovate faster for example.
The trap of buy vs. build
If the cost of building drops a lot, we might fail into the trap of custom-building solutions that could we just buy off the shelf. I think this might be happening in case of:
We already have teams in house, and we don’t (cannot) want to lay off people.
We need to decrease our spend, therefore we choose to stop paying for external tools.
We dedicate our teams to custom build solutions that we need to run our business because we are in control of where we spend, and the roadmap of those teams.
By the time you did that, you are adding unnecessary complexity to the organization to support those systems, and teams. Companies that didn’t go this direction will dedicate those teams to create more value add to their customers, vs. investing on reducing costs.
When everyone adopted the AI assistants, you will need to dedicate a lot of efforts, money, and energy into unique value proposition. If your teams are into improving costs instead of learning faster about the market, and customers, you will fall behind.
I'm still making my mind on the topic
This is a short post on a very dense topic, and we can only guess what will happen in the future. I would love to hear your thoughts on the matter, and if you want me to extend on some of the points I did.
Please, comment on how do you envision the AI assistants disrupting how we work and create products!
