In a brief dialogue, Kevin Surace explains how modern AI models like Claude can generate a functional Salesforce-like CRM in a weekend, stripping away the thousands of features of the legacy product. He argues that while startups can quickly build such bespoke systems to save on the high subscription fees, doing so shifts them from being users of a CRM to becoming software providers, inheriting governance, security, and database responsibilities. Surace notes that large enterprises are unlikely to abandon established vendors due to these risks, but a new class of low‑cost, AI‑built CRMs could undercut Salesforce’s pricing from the bottom up, creating a potential threat to the incumbent’s market. He concludes that the risk lies not in mass migration but in niche startups offering ultra‑cheap alternatives that could attract price‑sensitive customers.
Chapters
It took Salesforce $5 billion and years to build their product. Now 2 kids in a garage can copy it, and it only costs this.
No question that you can get Claude to code you essentially a lookalike Salesforce that has just the features that you need, not the 100,000 features they have, right? The downside is, are you really in the CRM business? No, you're manufacturing trinkets and selling them, whatever it is. You manufacture water pipes, you sell them, you use Salesforce as your CRM. So you're in the water pipe business, you're not in the CRM business. There is no question that you could code your own and bring it up, but now you have governance issues in databases, you have all kinds of things that they deal with, and security issues and all that, right? Now you're in the software business. I get that they're charging you $5 million a year, you know, to use Salesforce or whatever the number is. However, you're paying for a lot more than just the features in the software. So I suspect that while startups will indeed code their own CRM Because they want to save the $100K a year because it's real money they could put into something else. Large companies, which is the majority of business on a Salesforce or a Workday or a PeopleSoft or an SAP or something, are just not going to walk away from these vendors and say, 3 kids in a garage are going to code one up for me. They're just not going to. Now, here's the little fly in that ointment. Fly in the ointment is a startup in Silicon Valley, instead of spending the $5 billion that Salesforce spent to eventually get the product that they have today, could build Salesforce's product for nothing. And and take on the governance and the other things and sell their product for $2,000 a year. Just undercut, literally undercut Salesforce from the bottom up. And now you've got people who are paying $200,000 go, I'll just pay $2,000, I'll pay $5,000, right? So that's where I think the risk is. It isn't that every company is going to go build their own Salesforce. It's a ridiculous idea. But startups will build a Salesforce replacement and do so in a weekend. It used to take 1,000 Coders, 5 years. And now it takes 2 kids in a garage the weekend.
