Case studies & how we measure success.

Most SaaS websites start with testimonial logos. We don't have those yet — and we'd rather show you how we measure success than fake them. This page is the methodology framework: the metrics we track, what good looks like in production, and when first published case studies arrive.

Status as of April 25, 2026

First customer case studies: Q3 2026. We're in active onboarding with pilot dental, veterinary, and senior-living facilities through Q2 2026; published case studies will follow once we have at least 90 days of post-go-live operational data per case. We'll never publish synthetic or composite case studies — every case study you see here will be a real, named customer who's reviewed and approved the published numbers.

Until then, this page documents the framework we'll use, so you can evaluate Voipy on the methodology rather than the marketing.

The four metrics we track

For both Voipy Receptionist and Voipy Shield customers, we track four production metrics. Each is computed daily from your tenant's call data and surfaced in the customer-facing dashboard plus the optional weekly digest email.

Call coverage rate
covered / total inbound
Share of inbound calls Voipy answered before they hit voicemail. Receptionist target: 99%+ during business hours. Shield target: 100% (every call screens before reaching the resident).
False-positive rate
misclassified / total decisions
Calls Voipy classified as scam/spam that were actually legitimate. Receptionist sub-1% target; Shield sub-0.5% target on the auto-block confidence threshold (ambiguous calls route to staff review, not auto-block).
Scenario-to-booking conversion
booked / non-spam-eligible
For Receptionist: share of legitimate (non-spam) calls that resulted in a booked appointment, transfer to staff, or queued callback. Vertical-specific benchmark; reported per-vertical in case studies.
First-audio latency
p50, p95, p99 (ms)
Time from call connect to first AI audio. Production target: p95 <400ms warm cold-start. Reported on every case study because under-400ms is a hard differentiator vs vendors using cloud LLM APIs in the critical path.

What we publish per case study

Every published case study will include the same six fields, in this order:

  1. Customer name + vertical + plan tier. No anonymous "a Fortune 500 healthcare provider" framing. Real name, real industry, real plan tier — with the customer's signed-off permission to publish.
  2. Pre-Voipy baseline. What the front desk / call-screening process looked like before — captured during the trial-week measurement window.
  3. The four metrics (above) measured at day 30, day 60, and day 90 post go-live.
  4. Specific cost-side numbers. Hours of staff time reclaimed per week, dollar value at the customer's stated wage rate, and net plan cost — ROI shown not asserted.
  5. The customer's own quote on what changed in their operation. Edited only for length; never paraphrased into a marketing voice.
  6. What didn't work. Every case study includes a "what we'd do differently" section listing specific things the customer wishes Voipy did better. We publish negative feedback verbatim because case studies that only show wins aren't useful.

Why we won't fake testimonials before Q3 2026

Most SaaS sites at our stage either (a) display anonymized "stock" testimonials, (b) display logos of "trial users" as if they were paying customers, or (c) display synthetic quotes attributed to "operators" who don't actually exist. All three patterns convert better than no testimonials. We don't do them anyway, for two reasons:

What we publish in the meantime

The pre-customer-testimonial credibility surface relies on:

SurfaceWhat it documents
/scam-catalogThe full 93-pattern Voipy Shield library, each with its government-advisory citation. Auditable; you can verify each pattern against the cited FTC / FBI / SEC / FEMA / etc. source.
/securityCompliance posture: HIPAA BAA process, encryption specifics, 8-row sub-processor table, retention defaults, vulnerability disclosure with safe-harbor language.
/changelogPublic weekly release log. The library-update cadence is verifiable here, not just claimed.
/aboutFounder profile, parent-organization (SDN Bros) link, Q1 2025 → Q2 2026 timeline. No anonymous "founders" page.
/compare/12 side-by-side vendor comparisons. Every claim sourced.
/blog/5 published posts including 2 buyer's guides, 1 Q1 scam report, 1 engineering deep-dive on deepfake detection, and 1 dental playbook.

How to be a Q3 2026 case study

If you're considering Voipy and would also be willing to be a published case-study customer (with full editorial review of the published version), let us know on signup. Case-study customers get:

Email anton@voipy.app with subject "Case-study interest" and your vertical + plan tier you're considering.

What about the Voipy Shield family-facing report?

Shield customers have a separate quasi-case-study artifact: the weekly family digest. It's the plaintext email Shield sends to family members of protected residents listing what was blocked, by category, with dollar-loss estimates. Families forward these to siblings; the report functions as a self-contained case study of "what this product did for Grandma this week." We treat the digest itself as the de-facto Shield case study — every facility's family digest is real production data on real intercepted calls.

If you're evaluating Shield and want to see a sample anonymized weekly digest from a real facility (with names redacted, numbers preserved), email support@voipy.app.

Where to go next