This time last year was when I was first writing about what I perceived to be the need for a tool like PerformaGo. At that point, I was framing the problem in terms of lack of provision. L&D not being able to provide much in the way of post-training support. Primarily because it frequently wasn’t feasible from a resource or budget perspective.
My thinking then was that an AI-first performance support tool could cost-effectively and efficiently fill that gap.
The resource and budget issues remain as real and problematic as they did this time last year. But my thinking around the way that technology could fill that gap has both shifted and deepened.
And that shifting and deepening are most definitely connected to each other. Let me explain.
As I dug more into the need for a robust framework to underpin PerformaGo, I started to realise that I didn’t fully understand the entire performance support picture.
And I realised that was also true for most people in L&D, too.
The problem, I think, is connected to us confusing follow-on training with what I’m currently calling embedded learning.
L&D mostly provides preparatory training. It’s training that is there to get people ready for what they will be doing in the workplace. By its nature, it can only ever be representative of what learners will encounter back in the workplace. You can’t possibly cover all eventualities and how to deal with them.
If, miraculously, the resource and budget were to become available to provide some help post-training, we would almost certainly default to more preparatory training. Taking learners back into a formal training environment to cover all or some of the things we couldn’t cover initially.
Now, this isn’t a terrible idea but I don’t believe it is the optimal one, either. Because what I think it misses is that fact that, once back in the workplace, our learners become performers.
More follow-on, preparatory training takes them back into learner mode, when more than anything, they now need to be supported in performer mode.
They are still learning; but the nature of the learning has changed dramatically. It is now learning in situ – i.e., the embedded learning I referenced above.
A couple of things. Embedded learning will happen regardless of anything me might or might not do. It’s inevitable. But unsupported, embedded learning is uneven, messy, problematic and generally quite slow.
Effectively it’s learning by trial and error or, if you are lucky, trial and error and a more experienced colleague taking you under their wing and helping you along.
However, supported embedded learning can be a consistent, curated, highly-focused equivalent of that colleague taking you under their wing. This helps to even out the bumps, reduce the mess and crucially, speed up the process of developing more competence and expertise.
And it was that deeper realisation of what I believe is really required post-training that shifted my thinking. It moved me away from a pure AI-focus to what I’m now calling a performance support kit. Something that is is enhanced by AI but only in very specific cases.
Why this shift? There is, I think, a very fine line between providing ‘just more training’ and genuinely supporting embedded learning.
That kind of support needs to be carefully curated – either by a training-savvy SME or by L&D working closely with subject matter experts.
The key is asking the right questions to properly shape the nature of that embedded learning support.
Brilliant as AI is at many things, it’s still not that hot when it comes to extracting and curating exactly the right content from a subject matter expert to support a very specific performance need.
That is still very much a domain specific, human-first skill, based on years of hands-on workplace experience.
If you are interested, I have written about this difference between providing more training and supporting embedded learning in this piece on the Learning Re-Framed Substack.
Until next time…
