What are the gaps that the modern data stack promised to solve but is still unable to do so?

I think the biggest unfulfilled promise of the MDS is to make organizations do more with data. While this doesn't apply to every company adopting the MDS, a lot of medium-sized businesses, especially non-tech ones, haven't been able to jump on the MDS train due to the perceived complexity of it all.

MDS promised to be smarter, faster, and more modular. It promised to deliver better analytics. However, its yet to solve the problem of integration. Its still not extremely modular, and plug & play of different tools are not possible.Fragmented components in MDS, still make implementation and execution extremely difficult. As a result a lot of non technical smaller companies, fail to generate any value. Organizations still rely on legacy data stack for the ease of integration, setup and support when things go wrong.

One of the biggest problems that the modern data stack has yet to solve is true end to end data observability and lineage.  A new wave of data quality tools show a lot of promise in bridging this gap, but I have yet to see a stack that didn’t require a large amount of effort to guarantee a certain level of data quality and traceability. - Edited