Skip to content

Enterprise Bot achieves leading accuracy score on benchmark tests

At Enterprise Bot we have always focused on ensuring that our clients are able to get the best possible solution that will create wow moments for their customers. To achieve this the ability to truly understand what the user means is critical. That is why every year we invest heavily in achieving high accuracy in NLP and it is also the reason why we moved to more expensive, and resource extensive algorithms that leveraged bi-directional word vectors and deep learning frameworks compared to the market that was still on basic machine learning. We are happy to announce that our effort has helped us achieve the highest benchmark scores in the industry. For Benchmarking we have used two benchmarks

 

1. SNIPS Benchmark:

Here we were able to perform higher that all the previous scores by over 15% on average and scored higher on both the short and long versions of training data.

Enterprise Bot Benchmarking.png

 

2. Benchmarking Natural Language Understanding Services for Building Conversational Agents March 2021

This was a recent study conducted by Newsouth wales providing multiple frameworks and datasets to benchmark solutions. We are happy to announce that we have again beaten all industry solutions scoring higher on small and large dataset benchmarks. 

 

You can access the raw dataset here and our results and dataset here.

 

If you want to know how your enterprise can leverage the industry leading AI solution connects with us at sales@enterprisebot.ai or you can read more about our technology at enterprisebot.ai