Lighthouse AdvisorySLED AI Adoption Intelligence
← Back to results

From the Emergency Services edition of October 19, 2026

Independent reportingMixedNew this fortnight

Yadkin County, N.C. 911 says AI assistant AVA saved dispatchers hundreds of hours on its non-emergency line, with about 23% of calls passed to humans

Yadkin County Emergency Communications · 911 and emergency communications · Yadkin County, North Carolina, USA

Publisher
88.5 WFDD
Original publication
2026-10-19
Source retrieved
2026-10-20
Read original source

What happened

WFDD (88.5, Winston-Salem public radio) reported Oct 19, 2026 in its Carolina Curious series that Yadkin County Emergency Communications in North Carolina, the first 911 center in the state to deploy Aurelian's AVA voice assistant, has used it since July to answer the 10-digit non-emergency line. Operations Director Amanda Wilhoit says nearly three of every four calls to that line are information requests, which AVA now redirects or answers with links, and that the tool has saved dispatchers hundreds of hours; officials say it saved 3.4 telecommunicator hours in its first 24 hours. AVA costs roughly half of the about $66,000 a year one added staff member would cost, and it offers language translation. About 23% of AVA calls were passed to dispatchers because the caller asked for an operator or used words such as shooting, fire or ambulance; some callers get angry or repeat 'operator' until transferred. Real dispatchers still handle emergency calls. The story says Yadkin is one of more than 60 centers nationwide using AVA.

Why it matters

A small rural center (budgeted for three positions a shift) gives counties a cost comparison against one added telecommunicator and a transfer share to plan around, useful for centers that cannot hire their way out of concurrency.

Evidence and measured results

Public-radio feature quoting the center's operations director; operator-reported figures, no independent evaluation.

Limitations and uncertainty

Short public-radio feature built on the operator's account; hours saved and the 23% transfer share are not broken down by month or call type; no data on misroutes, abandoned calls, caller satisfaction or 911 answer times; cost figure is approximate; photo and the first-day figure are credited to the vendor.

Put this evidence to work

Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-10-19; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.

Sales

Role takeaway
Customer problem
Small 911 centers lose focus on emergencies when the non-emergency line keeps ringing during concurrent 911 calls.
Stakeholders
  • 911 director
  • County manager
  • Operations manager
  • Sheriff and fire chiefs served by the center
  • County finance
Discovery questions
  • What share of your non-emergency calls are information requests?
  • How often do 911 and admin calls arrive at the same time?
  • What would one more telecommunicator cost you per year?
  • How will you count hours saved?
Value hypothesis
An AI non-emergency assistant could relieve a small center's admin-line load at less than the cost of one position, if transfer and misroute rates stay acceptable.
Potential engagement
Non-emergency line assessment covering call mix, transfer-trigger testing and a 90-day measurement plan.
Unsupported claims
  • Faster 911 answer times
  • Reduced overtime or staffing
  • Caller satisfaction improvements
  • Accuracy of emergency keyword detection

Pre-sales engineering

Role takeaway
Fit
Voice AI front end for a county 10-digit non-emergency line with keyword and request-based transfer.
Architecture and integration
Vendor-hosted conversational agent on the admin trunk, warm transfer to dispatcher consoles, SMS link delivery and translation.
Prerequisites
  • Call-mix analysis
  • Documented answers and referral directory
  • Transfer keyword list
  • Console routing for transfers
Constraints
  • Must never answer 911 trunks
  • Rural callers may resist AI
  • Keyword triggers can miss indirect emergency language
Security
Define vendor retention of recordings and transcripts, access controls and handling of translated content.
Proof of value
  • Transfer share by reason, monthly
  • Sampled calls checked for missed emergencies
  • Hours-saved method documented
  • Abandoned-call rate before and after

Delivery

Role takeaway
Work
Deploy and operate an AI assistant on the non-emergency line with governed transfer rules and reporting.
Dependencies
  • Vendor configuration
  • Phone system routing
  • County IT
  • Public communications
Ownership
County emergency communications operations manager.
Skills and adoption
Supervisors review sample calls and adjust answers; telecommunicators handle transferred callers who may be frustrated.
Governance checkpoints
  • Transfer triggers tested before go-live
  • Public notice of AI use
  • Monthly transfer and complaint review
Acceptance criteria
  • Zero missed emergency transfers in sampled audits
  • Transfer share and reasons reported monthly
  • Documented hours-saved method approved by the director
Risks
  • Missed emergency phrased indirectly
  • Caller frustration and repeat calls
  • Overstated hours saved

Implementation considerations

Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.

Architecture and integration

What must connect, and where does the AI sit in the workflow?

Vendor-hosted voice AI on the 10-digit line with keyword and request-based transfer to dispatchers, informational links by text and translation; 911 trunks untouched.

Governance

Who approves, reviews and stays accountable for outcomes?

Transfer triggers (caller request, emergency keywords) are the core safety control and should be documented, tested and reported; agencies should define what 'hours saved' counts.

Security and privacy

What data, permissions and controls need testing?

Call recordings, transcripts and texted links pass through a vendor; retention, access and translation data handling are not described in the story.

Accessibility and workforce

Who is affected, and what skills or accommodations follow?

Translation may help non-English callers; telecommunicators keep emergencies, while callers uncomfortable with AI need a fast path to a person.

Procurement

What should contracts, pricing and exit terms secure?

Annual cost cited as about half of one $66,000 position; buyers should get per-call or flat pricing and reporting rights in writing.

Operating model

Which teams own the service once it runs?

AI absorbs routine information calls on the admin line; dispatchers handle transfers and all 911 calls.

Publication history

  1. 2026-10-19Emergency Services · Issue 435 resources
Read preserved resource versions (JSON)

Stable resource ID: wfdd-yadkin-county-911-ava-nonemergency-23pct-transferred-20261019