IMAIL

Intelligent e-mail answering service for eGovernment

Project duration: 1st January 2008 -  June 30th 2011

The IMAIL project develops and elaborates services in the field of e-administration that can rationalise communication between authorities and the public. IMAIL is a research project conducted by DSV and CSC (KTH Royal Institute of Technology) in collaboration with the Swedish Social Insurance Agency and Statistics Sweden. The objective of the project is to partially automate these authorities’ e-mail reply procedures.

The Swedish Social Insurance Agency receives approximately half a million e-mails every year. Roughly a quarter of the questions could be answered with an automatic reply. By sending automated replies to simple and frequently asked questions, officers can save time which can be spend on more complex matters. This would lead to better utilisation of officers’ expertise, in turn resulting in improved service to citizens and making officers’ work less monotonous.

What is eGovernment?

EGoverment is a collective term for the migration to e-services that has been under way for the past few years in the public sector. New technology can make the services available round the clock, improve service to citizens and increase cost-effectiveness. The objective is to simplify private citizens’ and companies’ contact with public authorities. Many services in eGoverment can be based on language technology.

Objectives and visions for the project

IMAIL’s overriding vision is to develop world-leading eGovernment services to facilitate efficient communication between public authorities and private citizens and companies.
A secondary objective of the project is to promote the development of semiautomatic text-based communication services in general, which can benefit both authorities and large companies.

Automatic e-mail reply

A few systems are currently used for automatic e-mail reply. For example, a person e-mailing a company or an authority may receive an autoreply with answers to frequently asked questions. If this e-mail does not provide the answer to the question the person asked for, he or she can send the question to be answered manually by a handling officer or customer service representative.

The IMAIL project develops systems that recognise text patterns in e-mails. This is carried out through clustering and text mining. Individual officers are unable to overview the vast number of questions and cases stored by public authorities. With the help of text mining, new, usable information can be produced from stored text. This new information can be used in automatic e-mail replies.

A key aspect of the IMAIL project is to examine how primary users (handling officers) and secondary users (members of the public) perceive and feel about the new technology.

Publications

Dalianis, H., J. Sjöbergh and E. Sneiders. 2011. Comparing Manual Text Patterns and Machine Learning for Classification of E-Mails for Automatic Answering by a Government Agency, 12th International Conference on Intelligent Text Processing and Computational Linguistics, CICLing 2011, February 20-26, Tokyo, Japan, In Computational Linguistics and Intelligent Text Processing, Lecture Notes in Computer Science, 2011, Volume 6609/2011, pp. 234-243, Springer-Verlag Berlin Heidelberg, DOI: 10.1007/978-3-642-19437-5_19.

Knutsson, O, T. Cerratto Pargman, H. Dalianis, M. Rosell and E. Sneiders. 2010. Increasing the efficiency and quality of e-mail communication in egovernment using language technology. In the Proceedings of Electronic Government and Electronic Participation - Joint Proceedings of Ongoing Research and Projects of IFIP EGOV and ePart 2010, J-L Chappelet, O. Glassey, M. Janssen, A. Macintosh, J. Scholl, E. Tambouris and M. A. Wimmer (eds), Lausanne, Switzerland, Trauner Verlag, August 29 - September 2, 2010, pp 45-52

Dalianis, H., M. Rosell and E. Sneiders. 2010. Clustering E-Mails for the Swedish Social Insurance Agency - What Part of the E-Mail Thread Gives the Best Quality? H. Loftsson, E. Rögnvaldsson, S. Helgadóttir (eds.): In the Proceedings of IceTAL 2010 (the 7th International Conference on Natural Language Processing, August 16-18, Reykjavik, Iceland), LNAI 6233, Springer Verlag, Berlin Heidelberg New York, pp 115-120

Dalianis, H., O. Knutsson and T.C. Pargman. 2009. Using human language technology to support the handling officers at the Swedish Social Insurance Agency. In the proceeding of the Workshop on Design & Evaluation of e-Government Applications and Service 2009, (DEGAS' 2009). August 24th 2009 - in conjunction with INTERACT 2009, Uppsala, Sweden.