Gr8 AI Solutions

Sofía is a bilingual (Spanish/English) AI voice receptionist for U.S. immigration law firms, with Muse Spark as the brain that understands every call. The problem is economic. Millions of people need immigration services, but thousands of firms chase the same leads, so customer acquisition cost keeps rising. A small firm pays for every call it generates, then loses it when nobody answers at 9 p.m., on a Sunday, or on a busy day when everyone is on the phone. To cut costs, firms move phone work to offshore remote staff, which is hard to hire (fluent English and Spanish, reliable workers) and exposes sensitive client data to people outside the firm and outside the US laws. What we built: 1. A voice agent (Retell) that answers 24/7 in Spanish, English or Spanglish, identifies new prospects, existing clients, live-show viewers, authorities and vendors, and routes each one. In office hours it warm-transfers so no call is missed; after hours it books a free pre-evaluation on Cal.com in the caller's own timezone. It never gives legal advice. 2. A post-call pipeline powered by Muse Spark 1.3. Each transcript goes to the Meta Model API with a strict JSON schema; Muse Spark returns 18 structured fields (caller type, outcome, detention, appointment, source, priority and more), a Spanish summary, an urgency level and a next action. 3. Deterministic guardrails: known facts (office hours, transfers) override the model; detention always outranks other outcomes. 4. Every call becomes a row in Google Sheets seconds after hanging up, with the recording link. Why it matters for the firm: - Every paid lead is answered, 24/7, at a fraction of a receptionist's cost. - The firm knows each caller's situation before calling back. - Matters the firm doesn't handle are filtered and can feed referral partnerships. - Lead sources (TikTok, Instagram, referrals) are captured on every call for campaigns. - No appointment is lost: after-hours calls are booked, in-hours calls are transferred. - Sensitive data stays in the firm's own systems instead of with offshore call centers or remote workers. Because Muse Spark's output is structured and reliable, every call can trigger actions in the CRMs and Case Management Systems firms already use (lead creation, case notes, tasks, follow-up reminders), fully connecting the phone line to how the firm tracks each client. This demo covers intake, but the same architecture scales to pre-evaluations, client service calls and follow-up calls across the firm. All names, numbers and caller data in the demo are fictional. The intake design draws on our team's experience running sales and intake at an immigration firm; the Muse Spark pipeline, server, integrations and updated voice flow were built during the hackathon. Team: Aidee Garza Garces, Luis Gerardo Garza Gonzalez, Juan Francisco Galvan Garza (Gr8 AI Solutions).