City of Tallinn
Connecting Tallinn citizens with city services through a WhatsApp reporting chatbot
The challenge
Tallinn citizens could report problems through existing municipal channels, but many people either did not know these services existed or chose not to use them. Some followed local issues through Facebook and neighbourhood group chats, while others searched for the correct city contact only after a problem had persisted. Previous unanswered reports, unclear responsibility between departments and complicated communication processes also reduced people’s motivation to engage.
The challenge was to design a single entry point for customer support that would make it easier for citizens to report problems, organise requests in a way that matched users’ expectations and connect every submission to the appropriate municipal workflow. The solution needed to be accessible, trustworthy and simple enough to use without requiring residents to learn another unfamiliar system or download a new application.
Methods
Desk Research, Content Analysis, Concept Mapping, Semi-Structured Interviews, Card Sorting, Information Architecture Mapping, Brainstorming, Prototyping, Backend Development
Team
Saara Vällik, Joseph Kalu, Yuliia Bulakh
Duration
3 Months
Research
The project combined the analysis of real citizen requests with interviews and card sorting to understand how Tallinn residents experienced municipal customer support. By exploring the types of problems people reported, the channels they preferred, and the reasons they sometimes chose not to act, the team identified both behavioural and structural barriers within the reporting process. These findings became the foundation for the information architecture and final chatbot solution.
→ Desk research & information architecture mapping
The existing customer support ecosystem was analysed alongside five anonymised citizen emails and 22 requests submitted through Anna Teada. Using content analysis and concept mapping, recurring communication patterns, user pain points, and the main categories of urban issues were identified. These findings informed the card-sorting activity and shaped the proposed information architecture.
→ Semi-structured interviews & hybrid card sorting
Six permanent Tallinn residents participated in semi-structured interviews exploring how they report urban issues and interact with existing municipal services. The interviews were followed by a hybrid card-sorting activity, which revealed users’ mental models and informed a clearer information architecture for the final solution.
→ Information architecture recommendations
The research revealed that users often struggled to classify urban problems consistently. To reduce confusion, the proposed information architecture introduced clearer categories and subcategories while helping citizens identify the responsible city department for each reported issue.
Key Insights
→ Many citizens rarely used official reporting services
Half of the participants had never used Anna Teada or similar reporting channels, while some were completely unfamiliar with the platform. Instead, many followed local issues through Facebook groups or neighbourhood chats without reporting them directly to the city.
→ Previous experiences influenced willingness to report
Unresolved issues and the lack of feedback reduced motivation to contact municipal services again. When citizens could not see whether their reports led to action, they gradually lost confidence in the system.
→ A single entry point for reporting urban issues
Participants preferred a single place where different types of urban problems could be reported instead of searching for the correct department or communication channel for each individual issue.
→ Simplicity mattered more than additional functionality
Most participants preferred a mobile-friendly website instead of downloading another application. They wanted a short, guided reporting process that required minimal effort and helped them submit the right information.
→ Guidance was valued, but human support remained important
Participants welcomed chatbot guidance for submitting reports but expected requests to be handled by a real person. Automation was valued for convenience, while human support remained essential for trust.
Design Solutions
The research showed that citizens wanted a simple and familiar way to report problems without searching for the right department or downloading a new app. Privacy, ethical, and trust concerns also made an AI-based solution unsuitable. These findings led to a WhatsApp chatbot that guides users through the reporting process and connects requests directly to the city’s Jira workflow.
→ Familiar and accessible reporting channel
Problem – Citizens did not want to download or learn how to use another application.
Solution – The reporting service was integrated into WhatsApp, allowing residents to communicate with the city through a platform they already used. Once the chatbot contact was saved, users could return to it directly without revisiting the Tallinn City website.
→ Guided issue reporting
Problem – Citizens were often unsure how to report a problem or what information the city required.
Solution – The chatbot guides users through a structured reporting process, collecting contact details, issue descriptions, location information, and optional images before submitting the request.
→ Automatic request management
Problem – Reported issues could require manual assignment before reaching the responsible employee.
Solution – Submitted requests were automatically transferred to Jira, where they were assigned to the appropriate staff member and managed through a shared workflow.
→ Clear access to customer support
Problem – Citizens needed a simple and visible way to start reporting a problem.
Solution – A persistent WhatsApp button was added to the Tallinn City website, providing direct access to the chatbot from any page.
→ A recognisable chatbot character
Problem – Automated customer support can feel impersonal and difficult to relate to.
Solution – The team created Killu, a chatbot character inspired by Tallinn City’s visual identity, giving the service a recognisable and approachable identity.
Usability Testing
The functional chatbot was evaluated through task-based usability testing to understand whether users could successfully report an urban problem, follow the conversation, and complete the reporting process with confidence. Participants completed the reporting task using the Think-Aloud method before sharing their feedback on the overall experience. They found the chatbot easy to use, appreciated the clear conversation flow, and valued being able to report issues through a familiar messaging platform.
The evaluation also identified opportunities for further improvement, including simplifying the initial interaction, making the opening messages clearer, and supporting automatic location sharing. These insights informed the final design iteration and helped create a smoother reporting experience.
Outcome
User-centred design helped create a simpler and more accessible way for citizens to communicate with municipal services. By combining user research, information architecture, development, and user validation, research findings were translated into Killu—a functional WhatsApp chatbot that streamlines problem reporting and integrates directly with the city’s customer support workflow.
The final outcome was a functional chatbot prototype, supported by a user research report and information architecture recommendations for Tallinn City’s future customer support services.