A scraper extracts data from various unstructured data sources such as PDF files and websites from towns and
cities. FAST leverages taxonomies and content from other municipalities in order to improve its
understanding of the content and context.
From unstructured content to knowledge
FAST uses state of the art natural language technology to extract knowledge about business processes from
this unstructured data and uses that knowledge to generate structured processes. These structured,
machine-readable processes can be used to further automate the business. A new standard - called OSLO-STEPS
- was created for this.
Automatic process optimisation
The generated processes are validated and further optimised: duplicated steps are removed, loops are removed,
irrelevant steps are removed,...
Automatic UI generation
A library of generic UI components is used to automatically generate a conversational UI out of the processes
that were the output of the previous steps. Both multi-step webforms, chatbots and even voice-based
conversational interfaces are supported.
Optimised UI based on users typology
The generated UI will be optimised for each specific user. Both the process itself and the UI are tailored to
the user. Based on what we already know from that user, we can skip certain steps to avoid asking the same
questions multiple times. The UI itself will be optimized based on the user’s typology in order to show
him/her the most user-friendly and easy-to-use UI.
FAST automatically generates different UI candidates based on the process definition. A/B testing is then
used to evaluate the different UI candidates and select the most optimal version based on actual end-user
behaviour and his/her typology.
Feedback loop for continuous improvement
Feedback from users is looped back into the previous steps in order to continuously improve the generated
processes. FAST will thus help governments to identify inefficiencies in their customer journeys and allow
them to further improve the overall citizen experience in dealing with government administration.