Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Hotline topic
No spam. Unsubscribe anytime.
Keystone previews AI-powered hotline to speed short-term rental complaints
Summary
Town clerk Maddie Selu showed a demo of a new AI-assisted STR hotline that will triage complaints, prompt users on whether to contact HOAs or emergency services, and route notifications to responsible agents via text, call or email; staff said citation actions will still require human follow-up.
Keystone town clerk Maddie Selu on Thursday previewed a new short-term rental (STR) complaint hotline that staff say will triage incoming reports, help callers identify the correct contact, and deliver faster notifications to responsible agents.
Selu said the town has been using a vendor, Decard Technologies, that “did not work quite like we anticipated,” and staff are developing a custom hotline interface with Blue Vector AI, a smaller local company. “One of the benefits of the AI system is it’s just able to help kind of triage it,” Selu said, describing prompts that will tell a caller that an issue may be an HOA matter or an emergency and recommend next steps.
Why it matters: Residents repeatedly told staff in listening sessions that they don’t know whom to call — HOA, police or the STR hotline — and that complaint reporting often arrives after an incident, making it hard to enforce the one-hour response requirement for responsible agents. The town also received repeated feedback that hotline interactions rarely closed the loop for complainants.
How it will work: Selu said the chatbot will ask whether a caller can provide a full street address, help match ambiguous location details to a legal unit, and flag high-priority calls (including suggesting callers call 911 where appropriate). Once a complaint is lodged, the system will notify the responsible agent by SMS, phone call and email and allow the agent to reply inside the platform. Complainants can accept the proposed resolution, reply to the agent, or escalate the matter to town staff.
“None of the complaint resolution or citation process is getting managed by this chatbot,” Selu said. Town staff will still handle formal citations and the follow-up work “on the following business day,” she added.
Costs and timeline: Selu said the town is paying about $3,000 a month in development costs now and expects the custom development phase to end; she said the town does not expect the same large ongoing licensing fees some vendors charge. Staff hope to launch the updated hotline in roughly six weeks, though implementation depends on staffing transitions: the clerk’s office is hiring a new town clerk and Selu said she will remain involved in the rollout.
Data and public records: Selu told the meeting that complaint lists published after the session will omit complainant names unless someone files a public-records request. She said the town will publish a complaints log and resolutions from the current season for community review.
What the public asked: Participants raised privacy and accuracy questions, and asked whether the AI could misclassify HOA-restricted behaviors as town violations. Selu said the AI’s prompts will be based on town ordinance text and staff will intervene when callers override suggestions.
Next steps: Staff will continue bi-weekly reviews with the existing provider while building the new system, publish the seasonal complaints/resolutions log, and finalize implementation after the new clerk is onboarded.
