Case Study: How an AI Voice Agent Captured 34 SIL Leads in 30 Days

An NDIS SIL housing provider in Sydney was facing a problem that's invisible in most dashboards: lead leakage after hours. Their intake coordinator handled calls during business hours — 9 to 5, Monday to Friday. But NDIS families don't operate on business hours. Support coordinators call during lunch breaks. Parents research SIL options at 10 PM after their adult child is settled. Guardians call at 7 AM before work.

The provider had voicemail. The callback rate was 40%. That means 60% of after-hours callers — families actively seeking SIL accommodation — never heard back because they'd already contacted another provider by the time the intake team returned calls the next morning.

The business impact: at an estimated 147 after-hours calls per month and 60% drop-off, the provider was losing approximately 88 potential enquiries monthly. With a typical SIL participant generating $200,000-$400,000 in annual plan funding, even capturing 2-3 additional move-ins per year represents a transformative revenue impact.

We deployed Riley — a custom AI voice agent built on the Vapi platform. Riley was configured with:

Challenge 1: Emotional callers. SIL enquiries are often emotionally charged — a parent worried about their adult child's housing. Riley needed to handle these conversations with appropriate empathy, not just clinical efficiency. We spent two full days refining Riley's conversation prompts to balance warmth with information gathering.

Challenge 2: Complex eligibility screening. Not every caller qualifies. Some have NDIS funding but no SIL component. Some have SIL but need a support level the provider can't accommodate. Riley needed to screen without making callers feel rejected — redirecting ineligible callers to relevant resources instead of dead-ending them.

Challenge 3: After-hours booking. The intake coordinator's calendar needed real-time availability checks. We integrated with Cal.com so Riley could offer specific appointment slots: "I can book you in for a property tour this Thursday at 2 PM — would that work?"

For the CEO: Three move-ins in month one. At conservative plan values, that's $600,000+ in annual funding from a $2,000/month AI agent investment. Occupancy across both properties moved from 78% to 92%.

For the intake coordinator: Monday mornings transformed. Instead of a voicemail backlog with half-remembered messages, she arrives to a dashboard of fully qualified lead profiles — name, contact, funding status, support level, preferred suburb, move-in timeline, and a call transcript she can review in 2 minutes.

For support coordinators: Response time to their referrals dropped from 24-48 hours to immediate. Two coordinators who had stopped referring re-engaged within the first month.

For families: Instant human-like response at the moment they're ready to enquire. No more leaving voicemails and hoping for a callback. Immediate SMS confirmation with next steps.

"Won't callers be upset they're talking to AI?" In 30 days of operation, fewer than 12% of callers requested a human transfer. Most didn't realise Riley was AI. Those who did commented that they appreciated the immediate response over leaving voicemail.

"What about NDIS compliance?" Riley doesn't make clinical decisions. She collects information and books appointments. All participant assessments, eligibility determinations, and onboarding decisions are made by qualified humans with the complete information Riley gathered.

"What's the actual monthly cost?" The AI agent platform, configuration, ongoing optimisation, and our management: $2,000/month. Against $600,000+ in captured annual funding, the ROI is measured in multiples, not percentages.

If we deployed Riley again from scratch, we'd integrate with the provider's participant management system from day one — currently, the intake coordinator manually transfers Riley's lead data into their system. We'd also add outbound follow-up calling for leads who submitted online enquiry forms but didn't book a tour.

Both are now in development for this client's Phase 2.