\n<\/p>\n
\n\t\tOpinions expressed by Entrepreneur contributors are their own.\t<\/p>\n
I\u00a0was talking to a founder who builds software for banks, and we drifted onto a theme I have not been able to let go of since: AI is aiding disintermediation and quietly removing vendor lock-in.<\/p>\n
If you think of old-school tech lock-in, perhaps the best example you can think of is banks and mainframes. The code is decades old, the vendor knows everything about it, and nobody on the inside wants to touch a migration. This is the founder\u2019s business. Between the contract lengths and the termination fees, he told me, \u201cIt\u2019s very, very sticky, and it\u2019s very, very hard to move away.\u201d<\/p>\n
That stickiness has been the business model for a big chunk of enterprise software. I think it is about to stop working, and the way I would frame it is as an opportunity for small businesses, in two different directions.<\/p>\n
Software vendors in particular assumed that if they served a piece of your technical infrastructure, they would be tricky to get rid of. And for most of the last few years, this has been true. If you provided someone with their core banking infrastructure, or the CRM system they use to manage their sales and customers and so on, well, it was risky to migrate away from you. Companies like Salesforce have managed to latch onto a huge part of the digital economy simply by being the vendor of choice and making it very difficult, risky and cumbersome to leave.<\/p>\n
Banking shows how extreme this gets. Reuters counted\u00a0220 billion lines of COBOL<\/a>\u00a0still in production, a 1959 language propping up 43% of banking systems. Leaving one of the big core providers, the founder told me, has meant \u201cdedicated teams to work through migrations over the course of 18 to 24+ months in order to move from one provider to the other.\u201d<\/p>\n