09-21-2017 Verena Fulde
- Chatbots to improve automated customer support, using the Twyla “Hybrid Brain”.
- A “conversation layer” for frequently complex enterprise data via seamless integrations with common enterprise systems.
- Already in the T-Systems Multimedia Solutions portfolio
The average time that customers are willing to wait for a customer support agent is two minutes, with impatient Americans at less than one minute, and some Germans accepting up to three minutes of horrible holding music. Unsurprisingly, the so-called Millennial generation is less patient than its older counterparts and also much more accepting of artificially intelligent chat technology.
So, it’s no surprise then that the recent explosion of artificial intelligence technology has found a warm reception from enterprise customer support departments. To be truly useful, however, such bots have to be well trained to provide the right answers, in the right context, and with the right tone of voice.
Twyla, the latest investment by Deutsche Telekom’s startup incubator hub:raum, has developed a “Hybrid Brain” chatbot solution that helps large companies to improve their digital customer care in a powerful and customizable way, and to rapidly address the distinct challenges faced by enterprise customer support. At a minimum, businesses can use Twyla to offload repeatable and trivial tasks to a trainable chatbot that can escalate to human agents as necessary. Further intelligence and scalability comes from Twyla’s ability to seamlessly integrate with the rich and frequently very complex ecosystem of enterprise databases, workflow systems, and software service providers on which large companies run their businesses.
“For the first time ever self-service content and conversational channels can be united using this emerging technology,” explains Paul Gibbins, co-founder and Managing Director of Twyla. “In this way we aim to help businesses to reduce customer support costs without negatively impacting the customer experience.”
The Hybrid Brain works by intelligently switching between machine learning and rule-based models and by dynamically converting rule-based structures into machine learning classifications once the TwylaBot has achieved a statistical confidence in what it wants to say to a user.
“This gives our clients all the power and scalability of machine learning, with all the control of rules,” says Gibbins.
Twyla was founded in 2015 by Gibbins and Chris Lotz and has grown to a team of 12, with more still to come this year. The two founders met at university in England back in 2007 and their careers took divergent paths before the two came together to found Twyla. Gibbins has 15 years’ experience in User Experience and Service Design with his career including roles at AOL and Burberry. Lotz built his career in management consulting, particularly in finance and most recently for Deutsche Bank. They were joined in the leadership of Twyla by Charles Yee, a PhD Computational Linguist, and CTO Ulas Turkmen who has a PhD in Cognitive Science.
Twyla is also backed by the founders of Berlin-based ProductsUp, as well as Frankfurt-based BoHa Capital.
T-Systems Multimedia Solutions, a subsidiary of Deutsche Telekom, has already created fruitful relationships between Twyla and its own client companies, seeking innovation and optimization in digital customer experience.
Email Newsletters
Sign up to receive TelecomTV's top news and videos, plus exclusive subscriber-only content direct to your inbox.