AI embedded in existing tools
Every software company is racing to embed AI into their existing products. Your CRM, your EHR, your accounting software, your project management tool — all of them are adding AI features. Some of this is genuinely useful (meeting transcription built into Zoom, for example). Some of it is checkbox marketing. The practical question for your organization is not "does our software have AI?" but "does the AI in our software actually solve a problem we have?" Often the answer is: the embedded AI is a starting point, but real operational intelligence requires connecting data across systems — which is exactly what those embedded features cannot do because they only see their own data.
Go deeper
Your EHR vendor just announced AI-powered clinical note summarization. Your project management tool now has an AI assistant. Your accounting software offers AI categorization. You're paying for all of it through bundled pricing whether you use it or not. Here's where it gets real: your intake coordinator is using the EHR's AI summary to brief clinicians on new patients, but the summary only sees what's inside that EHR. It has no idea the patient called your front desk three times last week with escalating concerns — that data lives in your phone system.
The trap most companies fall into is assuming embedded AI features eliminate the need for a data strategy. They don't. Each embedded AI sees only its own silo. The real operational intelligence — the kind that tells you a location is about to lose its best technician, or that a payer's denial rate just spiked — requires connecting signals across systems. Embedded AI is a feature. Connected intelligence is a strategy.
Questions to ask
- For each AI feature our vendors have shipped, is anyone on our team actually using it, and is it accurate enough to trust?
- Which of our critical operational questions require data from more than one system to answer?
- Are we paying AI surcharges on tools where we haven't even enabled the AI features?