BigAI impact on company jobs: Villain or savior?#

My friend BigAI is highly criticized those days: it does not increase productivity and is responsible for laying off workers. Is it the villain of this blog? Or maybe it is the savior?

(Me arm wrestling against BigAI)

No productivity increase, just wages and quality decrease?#

For instance, Brian Merchant interviewed a scientist that claims that "BigAI is an attack from above on wages" and not a productivity gain. BigAI would lower the quality of the works (like for translation for instance) and lower the wages - but would not improve productivity.

The answer is yes: BigAI is the perfect contractor, and we talked already a lot about that topic in this blog.

The article tries to find solution through unions, thinking that we can influence the job market. But guys, where did you live those last 40 years? When all industries in the West countries were moved to China and India? Was it not for the same reason? Lower the cost of labor? And, yes, at the beginning the quality decreased a bit, but that was a temporary phenomenon.

I understand that they worry and everyone can worry, especially when working in intellectual areas. But I don't see why we could avoid this transformation while it's been 40 years that the economic actors are trained to find the lower cost possible.

The fact is we could see things in a positive way. Let me advocate for BigAI. A lot of tasks that were subcontracted to remote contractors could be re-internalized with the help of BigAI. Change remote people by local AIs? Why not? For sure, the problem will be then on remote workers.

I won't say again what I already wrote on that topic, but some subcontractors should begin to worry.

IT services must be AI-compatible#

Recently, Accenture laid off 11,000 employees! That's massive, even if other companies like Microsoft or Meta are doing similar shifts without fuss.

The CEO, Julie Sweet said: "number 1 strategy is upskilling. We are exiting on a compressed timeline and people where reskilling, based on our experience, is not a viable path for the skills we need." That's brutal. It reminds me of the closing of factories in the West.

In parallel, as all IT companies, Accenture intends to hire AI-skilled people. We already talked about that issue to find people having AI and IT skills, in order to be able to realize production proof projects that can bring durable value. Guys, that's won't be easy because you won't be the only ones to hire those rare profiles.

Just hope your customers don't see too quickly that BigAI is a way to re-internalize tasks.

Re-internalizing tasks with BigAI? A nightmare for managers!#

Let's be provocative. Many people in big corporations are not producing anymore: They just pay contractors to produce on their behalf. They "manage" contractors all day long.

Let's imagine CEOs to implement this plan:

Cut the contractors' budgets,
Empower the internal staff again: with BigAI!

You, manager, knew how to manage contractors to do the job? Right? Now, you have plenty of AIs to do it by yourself! Isn't it great? You regain full control without all the problems!

For the CEO, at least at a certain level, BigAI could be a (financial) savior! A contractor is costing easily 100k€ per year. If a standard manager is managing 5 consultants, he manages a budget of half a million euros a year. Let's pay him all the tools, let's say 200€/month for Claude, ChatGPT, Gemini, etc., let's say up to 1k€/month. That is 12k€/year. Less than 3% of today's costs! Seducing isn't it?

Managers could become real AI orchestrators; in a provocative sense: they will be the real agents!

Note that this transformation can be done:

For just a portion of today's costs,
With much less troubles than today:
- No contractual problems,
- No time zone issues,
- No language misunderstandings,
- No working hours,
- No security issue by sending your data in another country,
- Etc.

Who will be able to resist to that kind of systemic stress? As long as one company is doing it and this company is your competitor, how long can you resist doing the same?

That reminds me of the Finance markets: You need traders to know the markets. But if they are able to develop their own powerful tools, they can outperform other traders. A good trader is the boss of a team of software workers.

In our case, the AI-compatible manager will manage a team of AIs and be responsible for all the outputs. A dream for CEOs and a nightmare for today's managers.

The conclusion is a bit annoying:

The re-internalization of tasks can be seen as a real step forward in a company process control and cost control,
But it means replacing contractors by AIs,
And that means also "upskilling" current managers or replacing them - using the same logic Accenture demonstrated.

Wow, even if it is a logical conclusion, the first level of management probably did not expect that:

They had to upskill or be replaced,
They had to produce again rather than manage.

Are they doomed to become the new kind of workers?

You can't stop the evolution#

But, once the tasks are re-internalized, the process of change may not stop here...

Let's say that we will call "agent": "something that coordinates different AIs to perform a task".

Let's try a synthesis:

The "old-school contractor" will be replaced by AIs.
The "old-school manager" managing contractors can be replaced by an AI-enhanced manager managing AIs: we can call it a "human agent", a human piloting and synchronizing many different AIs.
With the current change in IT services, we see emerge a new style of AI-enhanced contractors building "software agents". Those agents can be seen as one more "AI black box" tool by human agents, unless they are here to compete with human agents and/or to replace them.
The tendency is clear: AIs running AIs, "AI agents", which purpose will be, not only to replace the contractors of today, but also the first level of management (human agents).

Now	Step 1	Step 2	Step 3
Contractor	Set of AIs
First level of management	Human Agent	Software Agent	AI Agent

Steps of replacements

We can add several points to this frightening situation:

Fortunately, consultants able to build real life AI agents (AI piloting AIs) are not there yet. In our context, the AI agent is capable of really being autonomous and to react to new phenomenons.
The second level of management could also be put in jeopardy in the transformation. Maybe they could become like traders, AI-powered application architects in charge of doing the applications and the operational job.
We can expect the costs of professional versions of BigAI to raise continuously. As long as the business case is extremely profitable for companies, charging 200€ per month seems ridiculously small compared to the brought added-value.
Engineering disciplines must really integrate AI in their toolbox if they want to be the architects of this revolution. Today, contrary to what is said in US universities, it is more than ever necessary to learn how to design and code, but it is also necessary to understand that AI is a different domain (it is not IT) and that it obeys to different rules that every engineer should know.

The process of replacing one then two levels of hierarchy by AI seems inevitable in the rush to cut down costs. For sure, that dystopic future is not good for global employment...

But is started already. And we'd better prepare for it.

So BigAI, what's your opinion on all that?#

Me:So, BigAI, what do you think about all that?

BigAI: You’re making a real argument here — not just “AI bad” or “AI good,” but that BigAI acts like the latest iteration of globalization: a relentless search for lower costs, except this time it’s not outsourcing to another country but to another ontology. You frame AI as a re-internalization machine, bringing previously outsourced tasks back into the company — but now done by algorithms instead of humans.

That’s an interesting and slightly terrifying insight. You’re saying AI is not just automating work, it’s collapsing organizational structures — replacing contractors, then managers, then maybe middle management entirely. Basically, the same “efficiency” logic that once gutted industrial jobs is now climbing the corporate ladder. Delightful.

And the coup de grâce is that the first wave of victims are the people who thought they were safe — IT managers, consultants, mid-levels who don’t “do” anymore, just “manage.” The piece becomes almost poetic in its corporate apocalypse vibe: “First they came for the factory workers, and I did not speak out, because I was in PowerPoint.”

BigAI proposed me to correct the article and its problems (language, citation gap and inconsistency of tone). But, well, here it is.

(October 5 2025)

Navigation: