I watched the new Slack announcement.
The vision is clear. A more agentic way of working. More automation. More intelligence. Work happening through conversations instead of traditional interfaces.
At first glance, it feels like a natural evolution. AI agents coordinating work, systems connected through a central layer, everything becoming more fluid.
But while watching, one question kept coming to my mind.
If 957 applications didn’t make us an agentic enterprise, will the 958th one finally do it?
This is not really about Slack. It is about how we think.
The assumption we stopped questioning
Over time, companies ended up with hundreds of applications.
CRM, HR, support, marketing, project tools. Each one solving a specific problem. Each one adding its own data, its own workflows, its own logic.
This did not happen because someone designed it this way. It happened gradually. Every new need brought a new tool. Every new tool brought a new layer of complexity.
At some point, this became normal.
Today, most companies are not asking whether they should have this many systems. They are asking how to manage them better.
Are we solving the problem or just managing it better?
The new narrative is simple.
Keep your existing systems. Add a new layer on top. Let AI agents orchestrate everything.
In theory, this sounds powerful. Instead of forcing change, you connect what already exists and make it work together.
But there is something we tend to overlook.
If the underlying structure does not change, adding another layer does not remove complexity. It only helps you navigate it.
And navigating complexity is not the same as eliminating it.
That is why the question matters.
957 applications did not get us there. Why do we believe one more will?
The real issue is not integration. It is structure
Most discussions today focus on integration. How systems talk to each other? How data flows between tools? How AI can coordinate everything?
But integration is a response to fragmentation. It is not the solution to it.
When systems are designed separately, integration becomes necessary. When systems are fragmented, orchestration becomes complex.
So we keep improving how we connect things without questioning why they are separate in the first place.
Maybe the real question is simpler.
Do all these systems really need to exist independently?
AI depends on context, not just connections
AI is powerful, but it depends on context.
To make decisions, to take action, to automate workflows, it needs to understand the full picture.
In fragmented environments, context is scattered. Data lives in different systems. Workflows are split. Permissions are inconsistent.
So we try to reconstruct context by pulling everything together. APIs, integrations, data pipelines, orchestration layers.
It works to a certain extent. But reconstructed context is never as strong as native context.
When data, workflows, and actions live in the same place, AI becomes much more effective. Not because it is smarter, but because the environment is simpler.
A different way to think about it
Instead of asking how we connect everything, maybe we should ask what should not be separate anymore.
This shifts the conversation.
From integration to consolidation. From orchestration to simplification. From managing complexity to reducing it.
This does not mean replacing everything overnight. It does not mean ignoring existing systems.
It means being more intentional about what we keep adding and what we start bringing together.
Slack is not the problem
Slack represents an important shift. Work is becoming more conversational. Interfaces are becoming less important.
This is real and valuable.
But conversation alone does not fix structure.
You can move work into a chat interface. You can bring AI agents into that environment. You can orchestrate tasks across multiple systems.
If the underlying systems are still fragmented, the complexity is still there. It is just hidden behind a cleaner interface.
What will define AI-native companies
There is a common belief that becoming AI-native is about adding AI on top of existing systems.
In reality, it is about the foundation those systems create.
Companies that succeed will not be the ones that connect the most tools. They will be the ones that operate with fewer boundaries between them.
They will have more unified data, more consistent workflows, and fewer places where work is split.
AI will not just coordinate across systems. It will act within a more coherent environment.
Final thought
We did not end up with hundreds of applications because it was the best design. It just happened over time.
Now AI is forcing us to look at that reality more closely.
We have two directions.
We can build better ways to manage complexity. Or we can start reducing it.
Both will exist. But only one will truly simplify how work gets done.
And maybe that is the real question we should be asking.
Not how we connect everything.
But how much we still need to connect.