Remark has launched an AI chat agent to support cities manage 311 service requests. The new agent is designed to alleviate the backlog of 311 calls that cities across North America suffer during busy and critical times.

311 calls are non emergency service requests to city authorities. Each call can cost as much as $5.00 to resolve, and a lack of human agents limits its use.

Currently, 311 services receive millions of calls across its call centre, email, and mobile applications. Often complaint resolutions are not reported accurately or on time due to staffing problems or language difficulties. 311 operators are overwhelmed without sufficient access to the correct information or resources to provide accurate and timely assistance.

Using Remark AI’s proprietary large language model (“LLM”),  its Smart Chat Agent can communicate with callers and text requests using natural language to understand the context and urgency of the situation. Remark’s 311 agent simultaneously processes twenty service requests with LLM while supporting up to 40 languages.

By using machine learning and computer vision to see and verify an incident report. Accuracy is trained by a proprietary library of the client’s previous 311 transcripts and service requests. It can be fine-tuned and customised through relevant service experience.

The system automatically routes emergency calls to 911 channels, distinguishing a customer service request from vital emergency services.

“We are pleased to offer our AI-powered Natural Language Smart Chat Agent after numerous requests to help process and alleviate the backlog of 311 service requests faced by cities and townships around the United States and abroad,” said Kai-Shing Tao, Chief Executive Officer at Remark Holdings. “Too often, cities have to make the hard choice between spending to maintain citizen support services while allocating dollars during a declining budget environment. We are currently working with some of the largest cities in the world, who have come to us to help solve this crucial problem and we look forward to providing more details in the weeks ahead.”