Everyone I talk to in the Salesforce ecosystem has an Agentforce story right now. A POC that stalled. A demo that looked great in a sandbox and went sideways in production. A project that consumed months and delivered very little.
The narrative that follows is usually some version of “Agentforce isn’t ready.” I don’t think that’s the right diagnosis.
The implementations that are failing aren’t failing because of the technology. They’re failing because the data underneath it isn’t clean enough to support it. One number keeps coming up in my conversations with architects and technical leaders: only about 16% of global enterprises have data in good enough shape to run effective AI on top of it. That means roughly 84% of companies attempting Agentforce POCs are building on a foundation that was never going to hold.
This isn’t a new problem. It’s a debt that’s been accumulating for years, and AI just made it impossible to ignore.
Here’s what’s interesting though. The failures are starting to unlock something. Budgets for data cleanup, data governance, and integration work that couldn’t get approved two years ago are getting approved now, because leadership can finally see what bad data costs them in real time. The pain of a failed AI project is a more compelling argument than any technical debt conversation ever was.
If you’re a data architect, integration specialist, or data cloud practitioner, this is your moment. The work that positions companies to actually use AI isn’t the AI implementation itself. It’s everything that has to happen before the AI can do its job.
What are you seeing on the data quality side? Are your clients starting to invest, or are they still trying to skip that step?
If you’re trying to figure out where data and AI skills fit into your Salesforce hiring strategy, I’m happy to think through it with you. And if you need a Data Architect now but don’t have the time or appetite for a full hiring process, Oakcrest Talent Partners can place contract Data Architects through our employer of record, so you get the expertise without the overhead.
