VoiceAgent

Case Study: How Credit Unions Are Reactivating  Thousands of Dormant Accounts

Most dormant members haven't left — they've just gone quiet. These credit unions launched an AI voice agent that reaches them first, educates them, and brings them back.

february 25, 2026
Credit unions running dormancy outreach with AviaryAI

AUM: $6B+
Members: 360k+

AUM: $2.15B
Members: 57k+

AUM: $1.6B
Members: 115k

CDC Federal Credit Union

AUM: $437M
Members: 20k

THE CHALLENGE

You're Sitting on Thousands of Dormant Accounts. The Clock Is Ticking.

Every credit union is sitting on thousands of accounts that have gone quiet - members who haven't closed, haven't moved, just stopped engaging. And those accounts are drifting toward escheatment.

Letters go unopened. Emails get ignored. The staff member who used to make those calls has retired or been reassigned. Meanwhile, escheatment deadlines don't wait. Once funds transfer to the state, the member relationship turns adversarial - and that damage to trust is preventable.

THE SOLUTION

An AI Voice Agent That Reaches Every Dormant Member — Before the State Does

AviaryAI deploys an AI voice agent that works through your entire dormant account list — educating members about their account status, capturing reactivation interest, and routing engaged members to your team. In English and Spanish.

What changes for your team:

Every dormant member gets a call.

Not just the ones at the top of the list. The agent works the full queue on a schedule your team sets — thousands of accounts that would otherwise go untouched.

The conversation leads with relationship, not regulation.

Instead of "your account is dormant," the agent says "we haven't seen you in a while and wanted to check in before anything happens to your account." Member-first framing that aligns with credit union values.

You find out why members went quiet.

The agent captures reasons for inactivity at scale — login issues, loan payoffs, competitor switches, life changes. This is strategic data your team has never had access to before.

Interested members get routed to a human immediately.

Live transfers to a dedicated queue, Teams notifications with a 90-second accept window, scheduled callbacks, or branded follow-up emails with next steps. The human touch shows up exactly when the member is ready to act.

Spanish-speaking members are covered.

In some deployments, 25–50% of connected calls are in Spanish — coverage that would be nearly impossible to staff manually.

THE RESULTS

Results Across Live Deployments

Contact
Rate
28%
on members inactive for years
Reactivation
Lift
37%
vs. letters and emails alone
Engagement
Rate
67%
of connected members interact with the agent
AI Suspicion
Rate
<1%
members don't question it

Why Members Go Dormant

The insight that matters most:

Nearly 1 in 5 dormant members simply gave up trying to log in. That's not a retention problem — it's a fixable UX problem your team can act on today.

Why Dormancy First?

The Best First Voice Agent Use Case You Can Run

Dormancy outreach is where most credit unions start with AviaryAI — and for good reason. The volume is high (thousands of accounts), the risk is low (these members are already disengaged), and the success metric is clear (reactivation rate). Escheatment deadlines create a natural timeline, and the education-first approach aligns perfectly with credit union values. When this works — and it does — it opens the door to collections, onboarding, cross-sell, and every other use case your team has been thinking about.

IMPLEMENTATION - CRAWL, WALK, RUN

Launch in Two Weeks. Reach Every Dormant Member by Next Quarter.

Most teams start with a focused segment of 300–700 accounts approaching escheatment. From there, the agent expands coverage, adds channels, and begins feeding strategic data back to your team.

Phase
1
CRAWL

(Weeks 1–2)

Prove It Works

Launch on a defined segment — typically 300–700 accounts closest to escheatment. The agent calls after branch attempts have been made, leading with education-first messaging (not regulatory scare tactics).

Members who express interest get routed to a dedicated staff queue via live transfer, Teams notification, or scheduled callback. Every call is recorded with full transcripts for review.

Minimum data needed: First name, phone number, state.

You'll see: Connection rate, engagement rate, and reactivation interest within the first sprint.

Phase
2
WALK

(Weeks 3–6)

Expand Segments, Add Channels

Segment outreach by urgency (6 months–2 years vs. 2–3+ years dormant) and by account type or credit tier. Add email follow-up for members who express interest. Begin capturing reasons for inactivity as strategic data your team can act on.

Phase
3
RUN

(Weeks 6+)

Automate and Scale Across Use Cases

Integrate with core systems: SFTP call list feeds, Teams transfers, outcome writeback to your CRM or core platform. Set up recurring campaigns on a quarterly or ongoing cadence.

Expand beyond dormancy into adjacent use cases — 60/90-day inactivity outreach, promotional offers, collections, onboarding. Benchmark everything against control groups to measure true lift.

FAQ

Frequently Asked Questions

How fast can we launch?

Most teams go live in 2–4 weeks with a focused segment of accounts approaching escheatment.

What data do you need to get started?

At minimum: first name, phone number, and state. Most credit unions provide a filtered list from their core system.

Do members know they're talking to AI?

AI suspicion rates are consistently under 1%. Members focus on understanding their account status, not analyzing the voice.

Does the agent speak Spanish?

Yes. In some deployments, 25–50% of connected calls are conducted in Spanish.

How does the agent handle members who want to act immediately?

Live transfer to a dedicated queue, Teams notification with a 90-second accept window, scheduled callback, or branded follow-up email with next steps and disclosures.

What if a member asks to be removed from the call list?

The agent handles Do Not Call requests in real-time. DNC request rates have been under 1.5% across deployments.

Can we measure whether AI outreach actually works better than what we're doing now?

Yes. AviaryAI supports A/B testing against control groups. Across deployments, AI-targeted groups show 37% higher reactivation rates than groups receiving only traditional outreach.

See How It Works for Your Team

Live in two weeks. Thousands of members reached. No headcount required.