Kristoffer Fürst argues that managed integrations fail despite large catalogs. When a migration stalled because a vendor lacked a specific data transmission for a custodian, it exposed the core problem: some integrations to in-house systems can never be pre-built, so the entire model is broken.
Chapters
I was talking to a manager the other day and they were looking to set up a new integration from their investment management system. And they spoke to the vendor that they're using, they're not a client of ours, and that vendor has thousands of pre built integrations to various custodians, fund admins, etc. And they just assumed that this was going to be available. And then when it comes closer to actually changing custodian, which was the reason that they wanted a new integration, they realized that oh, this specific data transmission that we need to do for this custodian doesn't exist. They have other integrations for the same custodian, but not this specific one. And then they're dependent on the vendor creating that, so the whole project got delayed. And I think this is an example of an existing integration that could have existed. But there are also integrations that probably that can never exist, like an integration to an in house system, a data warehouse, etc. Those can never be pre built by the vendor. So the managed integrations, they're broken. It's not a good way to connect systems and it probably never was, definitely isn't today. So we need a new way to connect systems and service providers.