The Rural Support Service (RSS) is a public body of Latvia that is responsible for supporting agriculture, forestry and fisheries, supervising compliance with statutory acts, and performing other functions required by these sectors.
Nature of work that supports involving a chatbot
Unlike many other public bodies, the Rural Support Service provides more than 135 different services. So many services mean a large number of administrative procedures and extensive communication: providing information about aid, advice, processing and assessing requests, maintaining registers, and issuing various permits.
The nature of RSS clients’ business is also a serious challenge. Most often, it involves seasonal work, and their working hours often are longer that those of RSS consultants and regional offices giving customers limited opportunities to find the required information.
Information 24/7 and less work for employees required for simple issues (for example, by providing information on aid request deadlines) was the main reason behind RSS motivation to introduce a chatbot.
Virtual assistant Ieva joins the team of RSS
In 2020, other virtual assistants of public bodies was joined by Ieva — the RSS chatbot. Her task is to help clients to find answers to customer questions 24/7 both on RSS homepage and e-portal.
The chatbot is hosted on the joint platform of the Culture Information Systems Centre (CISC) and Tilde – Hugo.lv. It is a public administration language technology platform unique for the EU. On daily basis, the CISC maintains the virtual assistant, while Tilde is involved in implementing more complex integrations and functions.
When the chatbot was trained, there was a concern that it would be extremely difficult to teach Ieva as the RSS provides more than 135 services. Kristīne Ilgaža, Head of Public Relations of RSS, admits that the key to success was treating Ieva like a real individual and a colleague which requires specific knowledge, patience and persistence.”
Virtual Assistant Ieva’s dialogue scenario at the beginning of training (2020) and after 2 years
Stepping into the customers’ shoes
At the beginning, the FAQ of RSS customers were used to train the virtual assistant. However, as the SRR provides many services and processes large amounts of changing data, virtual assistant Ieva is trained and improved by a RSS chatbot trainer who teaches about various topics and updates its knowledge base on daily basis. For these purposes, the trainer uses questions that Ieva was not able to answer, website news, questions asked to Customer Service Representatives, and finds inspiration in social media: places where customers contact us every day.
Virtual Assistant Ieva’s performance and statistics in 2022
The personality of the chatbot is as important. As early as during the initial stages of the development, the image, communication style and language of the virtual assistant were defined, for example, how she addresses customers, what abbreviations and specific terms she uses. It is very important for the chatbot to avoid answering customers by reciting excerpts from laws or regulations, and to use user-friendly language.
“More than 4000 targeted discussions with a virtual customer service representative (chatbot) during one year may seem a lot to some, but not so much to others. The RSS believes that it is a lot, as the virtual assistant helps in serving customers that have simpler questions, while real RSS employees can provide explanations in more complicated cases. We trained Ieva, and also learned ourselves.
During these few years we realized that customers have accepted IEVA and do not want to trick it any more — instead, they ask serious questions. As our virtual assistant becomes more and more intelligent, we plan to involve IEVA in other activities, for example, to teach customers how to connect to the RSS portal or make changes in customer data. And this is just one example.”
Kristīne Ilgaža
Public Relations Manager of RSS
Ieva becomes more skilful
As the next steps, Ieva will learn to customize conversations and recognize voice messages. Customising conversations requires the ability to identify customers that will allow Ieva to solve targeted questions and rely on the personal data of the respective person ever more often. Moreover, voice recognition will allow our customers to contact Ieva via calls, and will enable RSS employees to record their observations during field visits to be automatically transcribed later.
Ieva has proven itself as a valuable support tool for customers by assisting with general issues online or in the e-portal, and by providing specific answers on how and where to apply for the support paid out by the RSS. Usage statistics of Ieva and chatbots from other public bodies demonstrates that the public is increasingly willing to rely on digital solutions offered by the public bodies.
