Dieses Stellenangebot ist nicht mehr verfügbar
Build Bilingual AI Phone System — Vapi + SmartSuite + + Twilio Integration
FreelanceJobs
- Canada
- Canada
Über
The Russel Moore Company is a commercial real estate development, brokerage, and acquisition firm based in Austin, Texas. We are seeking an experienced AI Automation Architect to build a complete "Virtual Office" phone and communication system for marketing commercial real estate properties in the McAllen, Texas market (Rio Grande Valley).
The system must handle all inbound and outbound communications through a single existing 956 area code vanity phone number, providing seamless bilingual support in English and Spanish, automated lead qualification, property information delivery, call routing to brokers, centralized data capture, email marketing, and owner-facing reporting dashboards. The entire system must be built on our existing SmartSuite and environment, integrated with Vapi (or comparable AI voice platform) and Twilio for telephony.
This is not a conceptual project. We have the vanity number, the SmartSuite workspace, and active property listings ready to go. We need a developer who can build, test, and launch this system within 6–10 weeks.
WHAT THE SYSTEM MUST DO
UNIFIED VANITY NUMBER INTEGRATION
We already have a 956 area code local phone number. This number will appear on all property signage, flyers, brochures, digital advertising, and email campaigns. The developer must integrate this existing number into the AI phone platform so that it supports voice calls (inbound and outbound), SMS and text messaging, and MMS for sending property images. The number must remain portable and under our ownership — it cannot be locked into any single vendor platform. All marketing materials will drive prospects to this one number, so it must function as the single front door for every interaction.
BILINGUAL AI VOICE AGENT — INBOUND CALL HANDLING
When a prospect calls the vanity number, they must be greeted by an AI voice agent that functions as a knowledgeable, professional front-office receptionist. The AI agent must perform the following:
Language Detection and Bilingual Support: The agent must automatically detect whether the caller is speaking English or Spanish within the first few seconds of the call and seamlessly continue the entire conversation in the detected language. This is not simple translation of scripts — the agent must be fully conversational and natural in both languages, able to handle questions, objections, and follow-up inquiries fluidly.
Property-Specific Knowledge Base: The AI agent must be loaded with detailed information about each listed property, including square footage, pricing, zoning, availability dates, location details, amenities, lease terms, CAM charges, and any other relevant property data. This information will be stored in SmartSuite and pulled dynamically so that when property details change, the AI agent's knowledge updates without requiring script rewrites. The agent must be able to answer common prospect questions without needing to transfer the call.
Lead Capture: During every call, the AI agent must collect the caller's full name, phone number, email address, company name (if applicable), which property they are inquiring about, the nature of their inquiry (purchase, lease, or investment), and their timeline. All captured data must be written directly into SmartSuite as a new lead record in real time.
Property Package Distribution: The AI agent must be able to automatically send property information packages to the caller during or immediately after the call. This includes PDF brochures, offering memorandums, site plans, and other marketing materials. Delivery should be via email and/or SMS based on the caller's preference. The system should trigger this automatically through workflows that pull the correct documents from SmartSuite based on the property of interest.
Call Notes and Summaries: After every call, the system must generate a detailed structured summary including who called, what property they asked about, what questions they had, their level of interest (hot, warm, cold), any specific requirements they mentioned, and any follow-up actions needed. These summaries must be stored in SmartSuite and optionally sent to the assigned broker via email or SMS notification.
SIMULTANEOUS CALL HANDLING (CONCURRENCY)
This is a marketing number that will appear on property signage in a high-traffic area. Multiple callers may call at the same time. The system must support high concurrency — multiple callers must each be greeted by independent AI agent instances simultaneously. No caller should ever receive a busy signal. The developer must architect the system to handle concurrent calls and should specify in their proposal how many simultaneous calls the proposed solution supports and what the cost implications are for scaling concurrency.
CALL ROUTING AND THE "ANTI-PERSONAL VOICEMAIL" PROTOCOL
Live Transfer with Warm Handoff: When a caller requests to speak with a specific broker, or when the AI agent determines that a live handoff is appropriate (for example, a highly qualified prospect ready to make an offer), the system must perform a "warm transfer" to the broker's mobile phone. Before connecting the caller to the broker, the system should provide an AI whisper briefing to the broker — a brief audio or text message telling the broker who is calling, which property they are asking about, and what they have discussed so far.
Voicemail Bypass ("Press 1 to Accept"): This is a critical requirement. When the system attempts to transfer a call to a broker's mobile phone, it must use "Press 1 to Accept" logic. The broker's phone rings, and when they answer, they hear a brief message such as "You have an incoming property inquiry about (Property Name). Press 1 to accept the call." If the broker presses 1, the call connects. If the broker does not answer, or if the call would otherwise roll to the broker's personal mobile carrier voicemail, the system must reclaim the call before it hits voicemail. The call should then return to the AI agent, which offers the caller alternatives: leave a message on the central system, schedule a callback, or receive the property package via email. Under no circumstances should a prospect's call ever land in a broker's personal voicemail inbox. All messages must reside on the central system.
The Callback Bridge: Brokers must be able to call into the central phone system from their personal mobile phones, authenticate (via recognized caller ID, PIN, or app login), and access a menu where they can listen to messages that have been left on the central server. After listening to a message, the broker should be able to press a button to automatically dial the prospect back. When the system places that outbound callback, the caller ID displayed to the prospect must show the 956 vanity number — not the broker's personal cell phone number. This ensures the prospect always sees and calls back the same vanity number.
Configurable Routing Rules: The system should support configurable routing rules including round-robin assignment between brokers, property-specific broker assignment (certain properties route to a specific broker), time-of-day routing (business hours vs. after hours), and caller preference routing (the caller says who they want to speak with).
If No Broker is Available: If no broker answers the warm transfer, the AI agent must seamlessly resume the conversation, offer to take a detailed message, schedule a callback appointment, or send the property package — ensuring no lead is ever lost.
OUTBOUND CALLING WITH VANITY NUMBER CALLER ID MASKING
This is a critical feature. Both designated brokers must be able to place outbound calls to prospects and clients while displaying the 956 vanity number as the outgoing caller ID — not their personal mobile phone numbers. The workflow should function as follows:
The broker dials into the phone system from their personal mobile phone (authenticated by their known caller ID or a PIN/app login). The broker enters or selects the number they want to call. The system places the outbound call to the prospect, displaying the 956 vanity number as the caller ID. When the prospect sees a missed call or wants to call back, they see and dial the vanity number, which routes back through the AI system.
This may be accomplished via a softphone app, a dial-in bridge number, or SIP-based calling from a mobile app. The developer should recommend the most reliable and user-friendly approach. All outbound calls placed through the system must be logged, recorded, and stored in the central communication archive in SmartSuite. The developer must confirm in their proposal that this caller ID masking approach is compliant with STIR/SHAKEN regulations and all applicable telecom rules.
CENTRALIZED COMMUNICATION SERVER AND ARCHIVE
All communications must be stored centrally in SmartSuite, regardless of whether they originated from the AI agent, from a broker making an outbound call, or from automated email campaigns. The archive must include:
Inbound and outbound call recordings with full transcriptions. All SMS and text message threads, both inbound and outbound. All emails sent and received through the system, including property packages, follow-ups, and mass marketing emails. AI-generated call notes and structured summaries. Lead contact records with complete interaction history showing every touchpoint. The archive must be searchable and filterable by date, property, broker, lead name, communication type, and lead status.
INTEGRATION AND AUTOMATION (SMARTSUITE AND MAKE.COM)
SmartSuite serves as the single source of truth for all property data, call logs, lead records, and reporting. serves as the integration and automation layer connecting the AI voice engine, Twilio, SmartSuite, and email systems. The developer must build all workflows in using webhooks, API calls, and scenario logic to ensure real-time data flow between all system components.
Specific automations required include: new lead creation in SmartSuite upon call completion, automated property package delivery triggered by AI agent or lead status change, broker notification workflows (SMS and email alerts for new leads and messages), lead status updates based on interaction history, and scheduled reporting workflows.
EMAIL MARKETING AND DRIP CAMPAIGNS
The system must include automated email marketing capabilities integrated through SmartSuite and
Mass Email Campaigns: Ability to send property announcements, price updates, availability changes, and new listing notifications to segmented prospect lists stored in SmartSuite.
Deliverability Tracking: Track email opens, click-throughs, replies, and bounces, with all metrics flowing back into SmartSuite dashboards so we can report on campaign performance.
Automated Drip Sequences: Trigger follow-up email sequences based on prospect behavior. For example, if a prospect opened an email but did not reply, send a follow-up three days later. If a prospect requested a brochure via the AI phone agent, enroll them in a nurture sequence. If a prospect clicked a link to view property photos, flag them as high interest and notify the assigned broker.
Branded Communications: All emails must be sent under the property or partnership brand identity, not individual broker email addresses. CAN-SPAM compliance with proper opt-out and unsubscribe management is required.
REPORTING AND ANALYTICS FOR PROPERTY OWNERS
The system must produce periodic summary reports — weekly, monthly, or on-demand — for the property owner or client. Reports must include:
Total number of inbound calls, texts, and emails received during the reporting period. Number of qualified leads generated. Summary of prospect inquiries including who called, what they asked, their stated level of interest, and their timeline. Email campaign performance metrics: number sent, open rate, click-through rate, reply rate, bounce rate. Outbound call activity by each broker. Follow-up status and pipeline progression for all active leads. Marketing activity timeline showing all touchpoints for the reporting period.
These reports should be generated automatically from SmartSuite dashboards and exportable as PDF for delivery to the property owner.
TECHNOLOGY STACK
The system must be built on the following platforms (already in use or selected):
SmartSuite — Central database, CRM, lead management, property data, and reporting dashboards.
— Workflow automation and integration layer between all system components.
(or comparable platform such as Retell AI or ) — AI voice agent for conversational call handling.
Twilio — Telephony backbone for voice, SMS, MMS, SIP trunking, caller ID masking, and A2P 10DLC compliance.
The developer should evaluate and recommend the optimal AI voice platform but must justify their recommendation with specific reasoning around bilingual capability, latency, natural conversation quality, warm transfer support, and cost per minute.
REQUIRED SPECIALTY SKILLS
Conversational AI: Deep experience with , Retell AI, or — specifically building production voice agents that handle real phone calls with natural conversation flow.
Database and CRM: SmartSuite expertise including table architecture, field relationships, dashboard design, and automation configuration.
Workflow Automation: Advanced scenario building including webhooks, API modules, routers, iterators, error handlers, and complex conditional logic.
Telephony: Twilio experience including SIP trunking, programmable voice, caller ID management, call recording, warm transfer logic, and A2P 10DLC compliance registration.
Bilingual Configuration: Native-level configuration for both English and Spanish AI agents — not just translation, but natural conversational fluency in both languages.
PROJECT PHASES AND TIMELINE
Phase 1 — Architecture and Compliance (Estimated 1–2 Weeks): Submit A2P 10DLC brand and campaign registration through Twilio. Map SmartSuite table architecture to the and Vapi workflow. Configure the 956 vanity number for simultaneous call handling and SIP integration.
Phase 2 — Build and Logic Development (Estimated 2–4 Weeks): Develop bilingual AI voice agent scripts and property knowledge base. Build the "Voicemail Bypass" (Press 1 to Accept) logic and the "Callback Bridge" system. Configure automated SMS and email property package delivery. Build email marketing workflows and drip sequences in Integrate all components with SmartSuite as the central data store.
Phase 3 — Testing and Scaling Preparation (Estimated 1–2 Weeks): Stress-test call concurrency with multiple simultaneous callers. Test all call flows end-to-end: inbound AI handling, warm transfer, voicemail bypass, callback bridge, outbound caller ID masking, bilingual language switching. Validate the "template" approach so the entire system can be easily cloned for future properties in different markets (for example, Austin, Texas) or for other business units (for example, car wash locations). End-to-end testing of email campaign delivery, tracking, and drip sequence triggers.
Phase 4 — Launch and Support (Estimated 1 Week plus 30 Days Post-Launch): Hands-on training for all designated brokers on how to use the callback bridge, place outbound calls through the system, access SmartSuite dashboards, review call logs, and generate owner reports. 30 days of post-launch monitoring, AI response optimization, script refinement based on real call performance, and bug fixes.
COST PROPOSAL REQUIREMENTS
Your proposal must include the following itemized breakdowns:
One-Time Setup Fee: Platform configuration, AI agent development, bilingual script writing, knowledge base creation, SmartSuite architecture, scenario building, Twilio configuration, testing, and broker training.
Monthly Recurring Costs: Itemize the estimated ongoing costs for Twilio voice minutes and SMS, Vapi AI usage (or alternative platform), SmartSuite subscription (if applicable), scenario credits, email sending costs, and any hosting or maintenance fees. We need to understand the monthly burn rate separate from the build cost.
Per-Use and Variable Costs: Per-minute AI call charges (broken down by component — hosting, STT, TTS, LLM, and telephony), per-SMS charges, per-email send costs.
Clone and Scale Pricing: Estimated hours and cost to replicate the system template for an additional property in a different market or for a different business unit such as car wash locations. This is important — we intend to scale this system.
Optional Add-Ons: Pricing for additional languages beyond English and Spanish, advanced analytics, expanded reporting, or additional concurrent SIP lines.
DATA OWNERSHIP AND INTELLECTUAL PROPERTY
All call recordings, transcriptions, lead data, contact records, SmartSuite configurations, scenarios, AI scripts, knowledge bases, and any other system configurations or data generated under this engagement shall remain the exclusive property of The Russel Moore Company. The consultant may not retain, reuse, or redistribute any proprietary data or custom configurations developed under this Scope of Work without prior written consent.
SCREENING QUESTIONS — PLEASE ANSWER IN YOUR PROPOSAL
How do you technically ensure a transferred call never drops into a broker's personal mobile carrier voicemail? Describe the specific mechanism you would implement.
What is your process for registering a brand and campaign for A2P 10DLC compliance to ensure SMS deliverability through Twilio?
Describe your experience using SmartSuite as a back-end database for voice AI call logs and lead management. Provide specific examples of SmartSuite architectures you have built.
How do you handle call concurrency to ensure multiple callers are answered simultaneously without a busy signal? What are the cost and configuration implications?
What is your approach to STIR/SHAKEN compliance for outbound caller ID masking on our existing 956 vanity number?
What is your approach to real-time language detection and bilingual conversation switching — do you use a single model or hand off between language-specific agents?
What is the expected latency on AI responses during a live call, and how natural does the conversation feel? Can you provide a demo or sample recording?
What happens if the AI cannot handle a question — what is the escalation and fallback process?
How quickly can the system be scaled to add additional properties or additional brokers to the collaboration?
Can you provide references from similar real estate, property management, or multi-user AI phone system implementations?
What is your recommended data retention policy, and how do you ensure all recordings and data remain under our ownership?
ABOUT US
The Russel Moore Company is a commercial real estate development, brokerage, and acquisition firm operating in Austin, Texas and the Rio Grande Valley. We manage a substantial real estate portfolio and operate multiple business units. We are technology-forward and already use SmartSuite, , and various AI tools in our daily operations. This is not our first automation project — we have experience with AI implementation and understand the technology stack. We are looking for a developer who can execute at a professional level, communicate clearly, meet deadlines, and build a system that is production-ready and scalable.
We prefer fixed-price proposals but will consider hourly engagements for the right candidate. The ideal candidate will be available to start within one to two weeks of contract execution.
Contract duration of 1 to 3 months. with 30 hours per week.
Mandatory skills: Twilio API
Sprachkenntnisse
- English
Hinweis für Nutzer
Dieses Stellenangebot wurde von einem unserer Partner veröffentlicht. Sie können das Originalangebot einsehen hier.