COBE the Chatbot – for Accenture, by Accenture
Accenture’s Legal, Marketing and CIO teams asked Fjord to help make Accenture’s Code of Business Ethics more user-friendly and, in so doing, showcase Accenture’s commitment to innovation and integrity in action.
For the first release, Fjord created an intuitive new website that makes it much easier to navigate the Code, embedded with a chatbot called COBE, and accompanied by more approachable content and simplified training as part of an integrated digital campaign. Release two is a more scalable model, with artificial intelligence and machine learning, enabling the chatbot to understand users’ context and develop its responses to their needs.
What we did
The challenge was to make required reading material about Accenture’s Code of Business Ethics more user-friendly, and to use it to project the ethical culture Accenture strives to create, both internally and to the wider world.
The company determined to invest in a new service for its own people that boasted the same excellent design standards and tech proficiency as products created for clients – all using its own in-house talents.
Collaborate and discover
Working alongside colleagues from across the Accenture ecosystem, we started out with a research phase to learn about how our employees need to consume the information in the Code, and to generate new ideas that would allow us to integrate it into our culture. Our aim was to turn the Code into something people actively wanted to engage with, rather than something they read because they are obligated to do so. We created an intuitive new website that is much easier to navigate, embedded with a chatbot called COBE, and accompanied by more approachable content and simplified training as part of an integrated digital campaign.
Approximately 17,500 Accenture employees worldwide interacted with COBE in its first four weeks after launch. The content team pulled together a huge, intricate matrix of potential keywords and relevant context, which fueled a decision tree. While the chatbot worked hard sifting through the matrix in the background, users’ experience was clean, efficient and relevant, enabling them to navigate their way through complex subject matter seamlessly.
For release 2, we leveraged user analysis and feedback to design a second version that wasn’t simply a tweak of what we had, but a complete regeneration. We’ve introduced artificial intelligence that has the capacity to interpret users’ context to ensure that the content it surfaces first isn’t just connected with their keywords, but accurately addresses their specific query. With machine learning built in too, every interaction with a user will be used to develop and hone the service.