In one of our recent blogs, Jayne introduced our Digital Voice programme, which uses the next generation of artificial intelligence (AI) to expand the services we provide to our customers. Barry Webb, who leads on the programme, talks about about how we’re building these capabilities and some of the outcomes we’re working towards.
One of my key priorities at the moment is leading on the Digital Voice programme, using leading edge conversational design techniques to develop automated services for our customers, delivered through AI, Robotics and Virtual Agents.
Over the past couple of years, Google has been utilising AI in a way which can be effectively used in a service environment, and have shared various videos
showing what their capabilities can deliver.
Here at BGL Tech, to support our delivery in effective design, we’re working with three companies who have been part of the Google experience in the UK: DVELP, Humanotics and Mazura. DVELP are a highly innovative and agile software development company, specialising in connecting emerging technologies like Google Contact Centre AI and legacy systems; Humanotics, led by Dr David Naylor who has a Ph.D in AI, are experts in the practical application of AI technologies; and Mazaru bring industry leading expertise in linguistics and conversational design.
Collectively, we have a great understanding of how to capture and learn from the ‘intents’, which we’re able to gather through all the data we have on our customers’ behaviour and the reasons behind their calls. This has been key in designing effective conversations to make it easy for our specific customers to use our services. We have, as we had understood and applied these techniques, captured some new design principles for Digital Voice design:
- Right interactions: Is the conversation intuitive? Just because you can automate something, doesn’t mean you should
- Design personas: Create intelligence, trustworthiness and likeability. Personas are critical to engagement
- Conversation design: Provide clear path. Users don’t make errors, designers do. Always act out scenarios to see what works and what doesn’t
- Multi-model: Use visual modes to supplement conversations. Design for use — use multiple modes only when necessary
- Test and iterate: Is the conversation intuitive? Real customer testing is critical — learn and adapt.
Using this model, we can ensure we’re utilising the latest technology to deliver first-class propositions for our customers. We’re starting to roll out pilots of our Digital Voice programme and I’m looking forward to seeing what our customer’s think!