Anne Hyland, VP of EHR Learning, Amplifire
I frequently hear some version of this question: once Epic finishes rolling out all of its AI and in-system features, will EHR training still matter?
I hear this from my colleagues, my clients, and mostly by my own sub-conscience! It’s a fair question. After hearing Epic present at the KLAS Arch Collaborative Learning Summit, and reading all of the news articles and LI posts from their annual UGM, there is no question that they are leveraging AI for improved efficiency across the healthcare ecosystem, and have a goal of making it “easier” to use with more in-system support features.
If the whole point of EHR training has been teaching people when, where, and how to click, and AI is increasingly doing the clicking, do we still need training? I think we do – but the needs shift. Here’s why training doesn’t go away.
- There’s a new competency – oversight and verification. The AI just sits on top of the core workflows – and clinician proficiency is still a core function, and the safety net. More importantly, AI is not perfect, and passive trust is a massive risk. For twenty years, EHR training has mostly been about execution: how to document, how to order, how to navigate a workflow without losing ten minutes to a wrong turn. That’s the layer AI is genuinely compressing. But somebody still has to verify, and that is not an execution skill, but a judgment skill, which is harder to teach. Clinicians need the judgment to catch even the most minor of errors, and where to look for them.
- The cadence of change is rapid. Adoption of AI is the fastest adoption healthcare has ever seen, and rolling out 150+ features over a year is faster than most (any?) orgs have managed or absorbed before. Increased change = increased need for training. AI tools mean new workflows, and new workflows mean new desired behaviors – behaviors that need to be learned, adopted, and used well.
- Adding AI doesn’t automatically translate to confident, efficient use. Historically, KLAS data has tied clinician satisfaction to training quality more than features. Adding capability without adding proportional training investment is exactly the pattern that produces a widening gap between what a system can do and what clinicians actually use well — the same gap that’s separated Elite-tier organizations from everyone else on ease-of-learning measures. AI doesn’t close that gap automatically. It’s just as capable of widening it, faster, if training doesn’t keep pace with rollout.
It’s safe to say that training will still be needed – but there will be a shift. What is being trained – AI oversight, prompt/output verification, exception handling, judgment calls – is different but is arguably a bigger need than documentation mechanics or order entry sequences.
In fact, my answer to the original question is that not only does EHR training still matter, but the training organizations that win the next 3 years are the ones that pivot content, invest in their learners, and ultimately gain even more from Epic than just new features before the next 150+ features ship.