U.S. Small Businesses Use AI to Build Custom Software Replacing Salesforce, Saving Over $100K Annually
0xBroomberg
At least five US small companies have ditched Salesforce or HubSpot contracts in the past six months, building replacements with AI tools and saving tens to hundreds of thousands of dollars a year — the first crack in traditional enterprise software's pricing power.
How are small companies actually saving money?
Atlanta real-estate firm Greenleaf Management (~55 employees) used Replit and Claude Code to build custom apps, replacing Salesforce and two other property-software vendors. Annual savings: ~$100,000. Monthly maintenance on the new system: just $300.
Seattle rugby team the Seawolves (70 people) had four engineers build a replacement with Claude Code in four months, dropping Salesforce CRM and ticketing platform AXS. Software spend fell ~$100,000; revenue since the season opened in March is up roughly 25% year-on-year.
Healthcare-software firm Hank AI (24 staff) and startup Atonom (45 staff) each walked away from $40,000/year Salesforce contracts. Their self-built replacements cost roughly $500 and $1,200 a year respectively.
This means → for companies with 20–70 people, AI coding tools have lowered the bar for "build your own CRM" from "hire a dev team" to "a few people, a few months."
Are large enterprises tempted too?
French pharma giant Sanofi (~75,000 employees) is cutting its use of ServiceNow, building AI agents with Claude Code and Cursor. The goal: shift 80% of workloads currently handled by ServiceNow, other software, and outsourcers — saving at least $10 million a year.
But Sanofi's chief digital officer admitted the push has met "considerable resistance" inside his own team. In plain terms = the bigger the company, the more custom workflows are baked into its systems — you can't just swap them out.
Enterprise travel firm Engine (~1,000 employees) has been a Salesforce customer for nearly a decade. A senior executive estimates migration would take at least a year; the company has no plans to leave.
This reflects a deeper source of stickiness: years of accumulated custom workflows — product catalogs, pricing rules, customer commitments — all wired into the existing platform.
Is the switch really worth it, or are hidden costs being underestimated?
IT consultancy Loka's CEO says a handful of clients have cut annual software costs by 40% to 80% by building AI replacements.
But he usually advises clients to keep existing SaaS, because the replacement process is slow and "pulls engineering resources away from things that actually differentiate the business."
His bluntest line: "No serious person is predicting the death of HubSpot or Salesforce." This means → the money saved on software licenses may come back as opportunity cost for the engineering team.
How are Salesforce and peers pushing back?
Salesforce's president told investors last fall that enterprise clients "have all tried to build it themselves," only to realize "you can't achieve enterprise-grade reliability and security with vibe coding."
ServiceNow cited a 97% contract renewal rate in Q1 this year, saying most customers are "expanding their use of ServiceNow, not leaving."
In plain terms = the incumbents' defense is simple: small companies leaving doesn't hurt; large customers can't leave. As long as renewal rates hold, Wall Street won't panic.
What is the real risk this points to?
OpenAI executives told investors earlier this year that the company's future products are expected to replace software from Salesforce, Workday, Adobe, and Atlassian.
Some large enterprises are increasing AI-vendor budgets while shortening contract terms with traditional software vendors or demanding better pricing.
This means → the core question is not how many small companies leave. It is whether improving AI coding capabilities will push the replacement trend from small businesses into mid-size and large accounts — if that transmission happens, SaaS giants face real erosion of pricing power over their biggest customers.
Content is for reference only, not financial advice.