Umwelt's Conversational AI, Nikki, helps to reduce employee attrition

Umwelt's Conversational AI, Nikki, helps to reduce employee attrition

I came across Umwelt.ai recently as a rather interesting application of Conversational AI internally.

I've done a lot of research on how Conversational AI can be tremendously effective serving both external customers and helping dramatically improve the efficiency of internal colleagues who are also serving customers.

I hadn't considered the possibility of applying Conversational AI to the internal human resources context. This is what Umwelt.ai offers.

This could be a really good use of the technology, depending on how it's implemented by clients – and the capability of the technology too.

I do like the idea of continuous and periodic engagement with employees – however, I wonder if there is a risk that given it's "a robot" rather than a human, there could be unexpected results. For example, I've seen many situations where by humans feel perfectly able to unload their innermost thoughts to a chatbot... especially one that's able to react in a ChatGPT-style caring and considerate manner. Employees could end up saying more than they mean.

Here's an example screenshot from the Umwelt.ai frontpage:

In the above example, I trust that these issues are then swiftly escalated – with due care – to the HR team for action. There's no point having fantastic technology in place listening to your employees if management/HR doesn't react or do anything as a result. Very quickly, employees will learn to ignore the technology.

I think it could be very effective though.

How much does it cost?

Well. There, I have a problem.

I have a special issue with any company that has a pricing page on their website – and then, when you click on it, the only option is, contact us.

Here's a screenshot from the Umwelt.ai Pricing page:

I don't think that's good enough in today's world, Team Umwelt. I really don't. Delete the page. Or put on some indicative pricing.

Anyway, I wanted to make sure I'd documented this use case!