How Winningtemp's AI helps you find trends and analyse the temperature

Winningtemp's unique employee engagement survey does not send random questions to you every week. The questions you receive depend on your responses to former surveys. 

Winningtemp's algorithm analyses response patterns continuously and asks questions based on relevance and variation. This means that your colleagues will probably get different questions every week. However, the algorithm only analyses responses, and there is absolutely no way of knowing which employee responded to which question.

For example: If you answer negatively to a question on stress one week, you may get a similar (or even the same) question the following week because the system wants to know if the responder was stressed just once or if it's a constant issue and needs to be addressed. 

The algorithms' findings are shown as Insights, which you can read more about here: What are Insights?


Are you interested to know more about our AI?

A few words about artificial intelligence (AI) from one of our in-house Data scientists.

"We can categorize AI as either general AI or narrow AI. General AI is a machine capable of learning to solve any number of problems without human input and would be able to adapt and evolve on its own.
On the other side, narrow AI is widely used. You could think of it as a "specialist" machine capable of handling a specific task or a limited range of tasks. These machines make autonomous decisions in their area of speciality and often outperform humans.

Winningtemp has several machines of this kind deployed in our product and some of them use machine learning (turnover predictions, hot topics). But what is AI, and how does it differ from Machine Learning (ML)? AI is anything capable of mimicking human behaviour to resolve problems in ways that are considered "smart". From the simplest applications to very robust and complex learning algorithms such as deep neural networks. Please note (also illustrated in the attached image), that in a broad sense, an AI / a machine can resolve problems in smart ways by using common statistical methods in combination with predefined rules.

You can think of the "Accuracy score" which given data and some predefined statistical parameters (confidence level 95%) calculates other statistical parameters (confidence interval), and if that calculated parameter is below a predefined threshold we say can be confident that the score will not change much. A more complex subset of AI are machines capable of learning, hence the name machine learning. These machines identify patterns in data, e.g., in past human answering behaviour, and can recognize similar behaviours and make decisions. They’re good at predicting, such as predicting employee turnover. They are getting better at their predictions every time they acquire new data. The only disadvantage is that humans often need to prepare the data by engineering what we think are important "data features" so that the machine has an easier job".

screely-1669625872534.png

Changelog 
2022-11-28 - Added more information about our AI

Articles in this section

Was this article helpful?
13 out of 13 found this helpful

Comments

0 comments

Please sign in to leave a comment.