Vaughn Ridley/Web Summit via Sportsfile via Getty Images
- AI coding tools like Bolt are disrupting the enterprise software landscape.AI lowers the barrier to building custom software, challenging traditional SaaS models.Companies may gain control and reduce costs by building in-house with AI tools.
The enterprise software landscape is being quietly, yet profoundly, disrupted by the rise of AI coding tools such as Bolt, Replit, and Cursor.
These services are upending one of the most fundamental technology decisions companies make: whether to buy software from external SaaS vendors or build it in-house. This once-clear distinction is blurring as AI lowers the barrier to building custom software.
I grasped the magnitude of this shift at a hackathon party in San Francisco. There, I met Netlify CEO Mathias Biilmann and the startup's security chief Mark Dorsi. They showed me a live Slack feed where new apps were being deployed on Netlify's platform, roughly one every 10 seconds. Many of these were created using Bolt and similar AI coding services. And most were internal tools built by companies for their own use.
This is a sea change. Historically, the decision to buy SaaS products stemmed from the high cost and complexity of building software internally. Skilled developers are expensive and internal IT teams are often stretched thin. But with AI-assisted software development, the economics shift.
Tools like Bolt empower a new class of developer — what Biilmann calls "AI-native developers" and what Bolt calls "software composers." These are non-traditional developers, often from business or operations roles, who can now build apps using AI to generate functional code through English language prompts. Training these new builders takes weeks or months, not years of expensive computer-science study.
As a result, there will likely be a huge influx of new developers. This new supply means businesses can hire more coders for less money. That, in turn, will help companies tackle software projects that previously wouldn't have made the cut.
"That's where the build-versus-buy equation starts changing," Biilmann told me. "When you have millions of new people who can build software, the barrier goes down. What a single internal developer can build inside a company increases dramatically. It's a much more radical change to the whole ecosystem than people think."
Where this shift will happen first
More than 10,000 new websites are being created via new AI coding tools and launched each day on Netlify's platform (versus being coded only by human developers), he noted.
Biilman sees the build-versus-buy shift happening in these software areas first:
- HR, Training, Q&A: These are often pretty simple applications that read and write data and do visualizations.
- Revenue Operations, CPQ, Business Dashboards: AI coding tools and agents can easily build interfaces or visualizations on top of existing company data.
- Marketing tools: These are relatively simple from a software perspective and they often need to be highly customized to specific company needs. What if AI agents write these applications?
"There are so many other areas that will be impacted by this, that it's daunting to think through the full implications," Biilmann added. "But these are some of the easy, low-hanging areas."
Who's building what
Netlify used AI coding tools to build an internal employee survey tool, a task typically outsourced to SaaS providers such as Qualtrics or Momentive.
Another example: a revenue operations staffer at Netlify used Bolt to create a pricing calculator for enterprise deals, potentially replacing a category of SaaS known as CPQ (Configure, Price, Quote) software.
And a Netlify recruiter built a new in-house Interview Training Course app for hiring managers with an AI coding tool called Lovable, instead of buying from an external provider.
Even VCs are embracing this trend by actually doing some of the work themselves. Martin Casado, a general partner at Andreessen Horowitz, recently built his own AI-powered customer-relationship-management tool. CRM software is the core offering of tech giant Salesforce.
"For some software, it's becoming quicker to code my own version with AI than learn someone else's non-intuitive, shit UI," Casado wrote on X recently.
Casado's CRM tool syncs with his calendar and email accounts. It takes that information and queries AI models to plan out each week, researching the people and companies he's due to meet.
A note of caution
Martín Migoya, CEO of tech consulting firm Globant, has spotted this trend.
"We're seeing projects that are coming out that way," he said in a recent interview, noting that some companies are beginning to try to replace unsatisfactory software services with internally-built alternatives.
Still, he urged caution. "Creating enterprise-class things is tough," Migoya added. AI models are improving, but software will still require support, the CEO noted.
His view reflects a common concern: maintenance and reliability. If something breaks in an AI-coded internal app, who fixes it? This is where infrastructure plays a critical role.
To address these concerns, companies like Netlify are developing what Biilmann calls "opinionated platforms" — infrastructure stacks optimized for AI-generated code and agents. These platforms handle authentication, authorization, staging, security, and data access in standardized ways, reducing the operational burden.
This evolution could enable companies to move from building prototypes to deploying production-grade software, without hiring an army of senior support engineers, he added.
The strategic implications for SaaS
The implications for the SaaS industry are significant. If companies can build tailored internal tools for the same cost, or less, than SaaS subscriptions, the traditional SaaS model could be under threat. Per-seat pricing may start to look expensive when compared to more abundant AI-assisted, in-house human coders.
Salesforce, the standard bearer of SaaS, could be vulnerable. While their core CRM system of record may remain sticky, some of the custom functionality layered on top might be replicated with AI-coding tools.
"Salesforce is really afraid," Biilmann said. "They are not going to be vibe coded away from being the system of record for all your sales data. But it's much more likely that a lot of their own custom functionality on top of that, that they charge extra for today, gets replaced by people building their own custom functionality."
Salesforce's answer, so far, is to build and launch its own array of AI agents. This Agentforce business is showing early signs of traction with customers, so this company may be well placed to fend off the threat.
Others may not be so lucky. Consulting firm AlixPartners recently warned that more than 100 midmarket software companies are stuck in the middle of this disruptive AI trend.
A massive new software layer
This new wave of software probably won't replace systems of record, it will sit on top of them. Think internal dashboards, pricing calculators, contract tools, HR survey apps, and media monitoring systems. All these are traditionally handled by SaaS vendors. But as internal developers gain new AI superpowers, building these tools in-house becomes a viable option.
AI-assisted software development is tipping the scales in the build-versus-buy debate. While it won't kill off the SaaS industry overnight, it introduces a powerful alternative.
Companies that embrace this model may be able to gain more control, reduce costs, and innovate faster. Those that cling solely to traditional SaaS may find themselves paying more for less, while their competitors build the future from within.
"You can now become a software developer without writing code," Biilmann said. "That's going to have a ripple effect on everything that's been built."
Sign up for BI's Tech Memo newsletter here. Reach out to me via email at abarr@busienssinsider.com.
