Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the AI And Digital Services topic
No spam. Unsubscribe anytime.
Arlington commissioners test county AI assistant AVA with senior-focused use cases
Summary
Commissioners and guests demonstrated three senior-oriented queries for Arlington’s AI assistant AVA — coordinating Uber with Metro Access, finding 55+ fee‑waivers, and a unified senior center calendar — and agreed to collect user feedback, involve service providers and clarify data and budget limits before expanding features.
Chair (S1) opened discussion about collecting civic use cases for AVA, Arlington’s county AI assistant, and invited a visitor from the Commission on Aging to present examples of how older residents might use the tool. The presentation focused on practical, layperson queries and whether AVA returned useful, actionable answers.
The guest (S4), who identified themselves as chair of the Commission on Aging, told members they tried three simple queries as a non‑technical user: “Can I use Uber with Metro?,” “How can I get a discount for 55 plus?,” and “What’s happening at the senior centers?” S4 said AVA’s answers were mixed — for paratransit the tool pointed users to Metro Access and the county’s Ability to Ride option that can use rideshare, but on discounts and consolidated activity calendars AVA either returned incomplete information or pointed users to multiple pages to click through.
Commissioners and staff said that result matched expectations for a first‑generation local assistant that draws primarily from county websites. Staff (S6) explained the system’s limits: “The data that we have in this case is the information that’s on the county website,” and AVA therefore cannot surface services or waivers that are not published in those sites. S6 also said the county’s vendor uses Microsoft’s service (TechGPT) and has incorporated Anthropic in parts of the stack.
Members emphasized next steps that do not require new technology: cataloging high‑value use cases, inviting service providers (Department of Human Services, Parks & Recreation and nonprofit partners) to supply representative prompts and data, and asking commission members and residents to trial AVA and report which queries failed and which succeeded. Staff proposed a staged, “playground” approach to pilot features and to estimate per‑query costs before wider rollouts.
The commission agreed to schedule focused outreach: invite the Digital Equity Coalition and relevant county departments to a future meeting, gather sample queries from frontline staff and service organizations (including Arlington Neighborhood Village), and compile user feedback for David’s technology team to evaluate. The presenters stressed the value of capturing user experience metrics — satisfaction, time to answer and whether a query required multiple page clicks — to guide budget and feature decisions.
The commission said it will continue the conversation at an upcoming meeting and expects staff to return with a plan for outreach, concrete use cases and a proposal for how to collect anonymized site search or AVA feedback data to inform improvements.
