July 31, 2026
Business Insider reported that Curative CEO Fred Turner canceled a $600,000-a-year Salesforce contract after his team built a replacement CRM in about two months, using AI to "vibecode" it themselves. That's not a pilot project or a proof of concept sitting in a sandbox. It's a company ripping out its system of record and replacing it with something it built in-house, on a timeline that would have been unthinkable for a CRM migration even two years ago. What makes this story land differently than the usual vibecoding demo is that Turner didn't stop at the CRM. According to Business Insider, Curative plans to cut roughly 80% of its SaaS spend this year and redirect that money toward AI. That's a company-wide bet, not a one-off engineering flex. When a CEO says yes without hesitation to the question of whether he believes in the "SaaSpocalypse," as Turner did, it's worth asking what he's actually seeing on his own books that gives him that confidence.
The detail that should get more attention than the Salesforce cancellation itself is Curative's Anthropic bill. Business Insider reported that spend grew from tens of thousands of dollars a month to millions of dollars a month over just six or seven months. That's not a company dabbling in AI tools on the side. That's a company reallocating what used to be SaaS license spend into inference spend, at a pace that mirrors exactly the kind of budget shift Turner is describing when he talks about cutting 80% of software costs. This is the pattern we've been tracking across the AI infrastructure market: money is moving toward the layer that actually does the work, not the layer that used to sit between a business and its data. We wrote about this shift in AI's Money Is Flowing to Infrastructure, Not Apps, and Curative is a case study in what that shift looks like from inside a single company's finance department.
The most concrete evidence in the whole story is Curative's AI agent, nicknamed Gwen, which handles healthcare contract negotiation. Business Insider reported that Gwen does this work for roughly $70 per contract, compared to $1,500 to $2,000 for a human doing the same job. That's not a marginal efficiency gain. That's a two-order-of-magnitude cost difference on a task that used to require a specialized, well-paid employee. When the economics on a single workflow look like that, it stops being a question of whether AI agents will replace parts of back-office and ops work and becomes a question of how fast finance teams can move budget from headcount and software licenses into agent infrastructure. Gwen isn't a chatbot bolted onto an existing tool. It's a purpose-built agent doing one job extremely well, which is a distinction that matters a lot once you compare it to how the big SaaS incumbents are responding.
Salesforce's response to this whole debate has been to push Agentforce, its own AI agent layer, and to argue the SaaSpocalypse fear is overblown. Salesforce CEO Marc Benioff has pointed out that Anthropic itself is still a Salesforce customer, and he has said Salesforce expects to close thousands of Agentforce deals in the fourth quarter. Those are real data points, and they're worth taking seriously. Anthropic using Salesforce today doesn't mean every company will keep doing so once its own team can vibecode a replacement in two months, and thousands of Agentforce deals closing doesn't tell you whether those customers are expanding their Salesforce footprint or just trying to squeeze more value out of a contract they're already locked into. That's the real issue with Agentforce as a strategy. It asks a business to add an AI agent on top of a CRM that's already expensive, already rigid, and already the thing generating the $600,000-a-year invoice Curative just canceled. You're paying to make the incumbent system smarter, not questioning whether you need the incumbent system at all. We've made this argument before in Buy vs. Build vs. ViibeStack: bolting AI onto a legacy platform doesn't remove the underlying cost or the underlying rigidity, it just adds a new subscription line on top of the old one.
The headline-grabbing part of the Curative story is the AI agent economics, but the more durable lesson is the two-month build timeline. Turner's team didn't need a large engineering org or a multi-year Salesforce implementation partner to replace a CRM that was central to how the business runs. They needed the ability to describe what they wanted and get a working system fast. That is exactly the gap ViibeStack's AI app builder is built to close, for teams that don't happen to have engineers willing to spend two months vibecoding a CRM from scratch. Our own CRM product, and the broader case we make on our page for teams looking to replace Salesforce, rest on the same premise Curative just proved out in public: you don't have to choose between an expensive, generic platform and an in-house engineering sprint. You can describe the sales pipeline, the contract workflow, or the support queue your business actually runs on, and get it built and deployed without hiring a development team. This also isn't limited to CRM. The same logic applies to the internal tools, ops dashboards, and workflow software that make up most of a company's SaaS stack, which is the argument we laid out in What makes an internal tool "purpose-built" vs. generic" and in our practical build vs. buy framework.
If you're currently deciding whether to add Agentforce to an existing Salesforce contract, the Curative story is a useful gut check. Ask what you're actually paying for. If it's a smarter layer on top of a platform you were already unhappy with, you're solving the wrong problem, and you're doing it at Salesforce's price point, not Curative's $70-per-contract price point. Benioff is right that not every company is going to vibecode its way off Salesforce this quarter, and a lot of large enterprises have integration commitments that make a two-month rebuild unrealistic today. But the direction is clear, and it's the same direction we've been pointing at all year: the businesses that come out ahead won't be the ones that layer AI onto their existing SaaS bill, they'll be the ones that use AI to stop needing that bill in the first place. That's the bet Curative made, and it's the same bet ViibeStack is built around.
Sources