For two decades, buying enterprise software meant accepting a vendor’s feature set, paying per seat and hiring specialists to manage the platform. For small businesses, that model often meant paying for capabilities they never used. AI coding tools are changing that calculation.
Five startups and small companies with staff ranging from 20 to 70 people ended contracts with Salesforce and HubSpot over the past six months in favor of applications they built themselves using AI tools from Anthropic, Lovable and Replit, The Information reported. The firms cut their software costs by 40% to 80%.
Research and advisory firm Gartner found that $234 billion of enterprise application software spending is exposed to agentic arbitrage by 2030, or roughly 20% of all enterprise SaaS spending. “Agentic AI changes the economics of software,” said George Brocklehurst, a managing vice president at Gartner, in a July 1 press release.
Retool, a low-code platform for building custom internal tools, found that 35% of enterprises have already replaced at least one SaaS tool with a custom-built alternative, with 78% planning to build more this year.
The Math Shifts When Software Costs $1,200 a Year Instead of $40,000
Greenleaf Management, an Atlanta-based real estate investment manager with about 55 employees, replaced Salesforce with a custom application built using Replit and Claude Code that costs roughly $300 per month to maintain, saving around $100,000 annually, partner Dave Codrea told The Information. The new app also let Greenleaf exit contracts with real estate software companies Entrata and Yardi.
Atonom, a 45-person Utah startup, made the same decision, replacing a $40,000 Salesforce contract with a Lovable-built CRM expected to cost $1,200 per year. “No one was using Salesforce to its fullest potential,” Atonom Chief Revenue Officer Gabe Larsen told The Information. “I don’t know if we need every bell and whistle known to man on CRM.”
The Seattle Seawolves, a 70-person professional rugby organization, used Claude Code to replace both its Salesforce CRM and AXS ticketing system in four months. Owner Adrian Balfour told The Information that the club reduced software spending by around $100,000 and that revenue is up 25% since the start of the season in March.
The shift is not confined to small companies. Sanofi, a French pharmaceutical firm with around 75,000 employees, is cutting 80% of its ServiceNow usage by routing work through AI agents built with Claude Code and Cursor, targeting savings of at least $10 million annually, The Information reported.
Vendors Meter AI Agents and Rethink Pricing
Enterprise software vendors are not sitting still. ServiceNow, SAP and Workday introduced controls requiring external AI agents to pass through metered integration layers to access customer data, PYMNTS reported.
Salesforce and ServiceNow are moving toward outcome-based pricing, PYMNTS reported. Salesforce’s AgentForce annual recurring revenue (ARR) grew 169% in a single quarter, according to Forbes.
Bobby Mukherjee, CEO of Loka, an IT consulting firm, told The Information that replacing SaaS platforms “pulls engineering attention away from things that actually differentiate the business.” Salesforce President Srini Tallapragada told investors companies “all try to do the do-it-yourself, and they’re realizing that you can’t vibe code your way to enterprise reliability and security.”
ServiceNow reported a 97% renewal rate in the first quarter of 2026, citing customers expanding their footprint rather than leaving, according to GuruFocus.
There is also a structural case for staying. What makes it difficult to move off Salesforce or any other SaaS provider are the layers of customized workflows companies continually add and update to track everything from product catalogs to pricing to customer commitments.
Enterprise SaaS platforms are also built to support compliance requirements, security infrastructure and integrations across dozens of systems that custom-built tools typically lack, industry blog The SaaS CFO noted.
Enterprise software spending is on pace to rise 15% to $1.4 trillion in 2026, Forbes reported. The money is not leaving enterprise software. It is moving to different products at different prices.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
