Christoph Börner, CEO, Botium, joins us on our podcast to talk about quality assurance for conversational AI systems.
Presented by Deepgram
Deepgram is a Speech Company whose goal is to have every voice heard and understood. We have revolutionized speech-to-text (STT) with an End-to-End Deep Learning platform. This AI architectural advantage means you don’t have to compromise on speed, accuracy, scalability, or cost to build the next big idea in voice. Our easy-to-use SDKs and APIs allow developers to quickly test and embed our STT solution into their voice products. For more information, visit:
Quality assurance for digital assistants
One thing is for sure; conversational AI systems are a challenge to test and maintain. Yet, without robust testing, there’s no way of guaranteeing quality. And high quality interactions are all that matter. Without assuring the quality of your assistant, it won’t achieve the results you hope it will, rendering your whole strategy pointless.
To share the importance of quality assurance, and to walk you through the core steps and activities in your QA pipeline, we’re joined by Christoph Börner, CEO, Botium.
00:00 Introduction and presenting Deepgram
Register for ‘The irony of big tech: why your speech recognition is probably a load of tosh’: https://vux.la/3ySe8n8
04:30 Welcome Christoph Börner
07:04 What is Botium?
09:20 Botium founding story
13:22 Challenges with conversational AI quality control
18:32 Benchmarking quality and testing
22:10 Is there privacy hurdles to overcome?
24:30 Gaps in AI practitioner knowledge
28:37 Removing complexity from quality control
31:17 Common issues found while training chatbots
35:00 Large language models
37:21 Training data best practice
40:00 Best practice for NLU selection
47:05 Monitoring the performance of digital assistants
52:55 Automated QA: the ultimate insurance policy
55:10 Team maturity