Here’s the second of three posts about last week’s Learning Technologies Exhibition and Conference and it’s a reflection on Markus Bernhardt’s session that I was lucky enough to chair.

Agentwashing was the word Markus Bernhardt used to open his session – not AI agents but agentwashing.

His argument was quite precise. Most organisations cannot tell you whether what they are calling an agent is actually an agent, or a chatbot wrapper with a new job title. As Markus said, if it doesn’t have agency, it’s not an agent. So we end up with dashboards that are full of data and green but these instruments are measuring the wrong thing.

An audience member asked why people are naming tools as agents when they’re not. My answer was simple: Agent is to AI what leadership development is to management training. Similar content, delivered differently with a nicer wrapper and a 20% increase in price.

Because if it’s true some of the time, the question Markus was really asking is: does your organisation have the diagnostic capability to tell the difference? Not the vendor’s word for it, not the dashboard, but your own judgement, based on evidence.

The shift Markus is asking L&D to make is from course designer to system designer. This is the entirely consistent with the position I’ve been arguing for over a decade – the move from shopkeeper to engineer. From measuring completion to measuring whether the system is working. That is not a skills gap but a role gap and many L&D job roles haven’t caught up. Yet.

Rachel Burnham captured the whole session in a single sketchnote and is worth seeing if you missed the session.

If you want to go deeper, find Markus Bernhardt and the Endeavor Report.