AI May Not Kill Software. It May Kill Software Categories.
One of the most common assumptions about AI is that software companies will disappear because software will become easier to build.
I think that's looking at the problem the wrong way.
The real risk isn't that AI eliminates software.
The real risk is that AI eliminates the need for certain categories of software.
In other words, some products may not get replaced by competitors. They may get absorbed into systems businesses already own.
Here are a few categories I think investors should be watching closely.
Salesforce: Does CRM Remain a Product?
Salesforce is one of the most successful software companies ever built.
The bull case is obvious.
The platform is deeply embedded inside thousands of organizations. It has a massive ecosystem, countless integrations, and significant switching costs.
But AI introduces a different question.
How much of Salesforce's value comes from owning a business process, and how much comes from providing functionality?
Many organizations use only a small percentage of Salesforce's capabilities. At its core, many deployments boil down to:
- Accounts
- Contacts
- Opportunities
- Activities
- Reports
- Workflow automation
Historically, building these systems internally was expensive.
Today, AI is dramatically reducing the cost of creating and maintaining custom software.
At the same time, CRM functionality can increasingly be embedded inside customer portals, ERP systems, support platforms, and internal operational tools.
The risk is not that another CRM beats Salesforce.
The risk is that CRM stops being a standalone destination.
HubSpot: When Simplicity Becomes a Liability
HubSpot built its business by making marketing, sales, and customer management accessible to small and medium-sized businesses.
That simplicity was a strength.
It may also become a vulnerability.
Small businesses are often the first to question recurring software costs. They are also increasingly able to use AI tools to build lightweight internal solutions.
HubSpot still benefits from strong brand recognition and an integrated platform.
But investors should ask whether customers are paying for ownership of a business function or simply renting functionality.
If the answer is functionality, pricing pressure could increase significantly over time.
Atlassian: Is Project Management a Product or a Feature?
Atlassian's Jira has become the default issue-tracking platform for much of the software industry.
The challenge is that issue tracking may be moving toward feature status.
AI can create tickets automatically.
AI can summarize work automatically.
AI can coordinate workflows automatically.
The question is whether dedicated project management platforms remain central destinations or whether they become embedded capabilities inside broader engineering and operational systems.
Atlassian's moat remains substantial today.
The long-term question is whether the category itself remains as valuable as it has historically been.
Wix and Squarespace: The Most Direct Threat
Website builders may face the most obvious disruption.
For years their value proposition was straightforward:
Building websites is hard. We make it easy.
AI is steadily making website creation easier.
At the same time, AI is reducing barriers to entry for competitors.
This creates pressure from both directions.
Customers can build more themselves.
Competitors can build more easily.
When both sides of the market gain leverage, the economics of the category become less attractive.
That does not mean Wix or Squarespace disappear.
It does mean investors should question whether the category's best growth years are still ahead.
Why ServiceNow, ADP, and Workday Look Different
To understand the difference, consider companies such as ServiceNow, ADP, and Workday.
These businesses are not primarily selling features.
They are selling ownership of business functions.
Companies do not buy ADP because they enjoy payroll software.
They buy ADP because they do not want payroll to be their problem.
Companies do not buy ServiceNow because they enjoy workflow software.
They buy ServiceNow because enterprise operations are complex and expensive to coordinate.
The distinction matters.
AI can make these companies more efficient.
It does not eliminate the need for someone to own the problem.
The Question Investors Should Ask
For years, investors have evaluated software companies by asking:
How difficult is this software to build?
In an AI world, a more important question may be:
If this company disappeared tomorrow, would customers replace it with another product—or would they simply absorb the functionality elsewhere?
The answer to that question may determine which software companies thrive over the next decade and which categories gradually fade into the background.
Because AI may not kill software.
But it may kill the idea that every feature deserves its own company.