Aventora Engagement Hub - Features

Aventora Engagement Hub is an enterprise-grade AI-powered phone automation platform that transforms how businesses communicate with customers through intelligent voice calls, SMS, and WhatsApp messaging. The platform enables natural, human-like conversations at scale, handling everything from appointment bookings and confirmations to customer support and information delivery. Whether you need to reduce no-show rates, automate customer outreach, or provide 24/7 support, Aventora Engagement Hub delivers professional, context-aware interactions that enhance customer satisfaction while reducing operational costs.

Executive Overview

Aventora Engagement Hub

What It Is

Aventora Engagement Hub is an AI-powered, multi-channel customer engagement platform that enables businesses to handle phone calls, SMS, and WhatsApp conversations automatically—using natural, human-like AI conversations that can take real actions.

Unlike traditional IVR systems, chatbots, or call-center automation tools, Aventora is designed to understand intent, hold real conversations, and complete business workflows such as booking appointments, confirming schedules, delivering information, and seamlessly handing off to human staff when needed.

Why Businesses Need This Now

Customer expectations have changed. People expect:

Immediate responses
Natural conversations
Availability outside business hours
Communication in their preferred language and channel

At the same time, businesses face:

Rising staffing costs
Missed calls and lost opportunities
High no-show rates
Fragmented tools for phone, SMS, and messaging
Compliance and data-privacy constraints

Aventora bridges this gap by acting as a 24/7 AI engagement layer that works alongside your team—not replacing them, but removing repetitive, time-consuming interactions.

What Aventora Does

Aventora engages customers across multiple channels and handles conversations end-to-end:

Answers inbound and outbound phone calls
Sends and receives SMS and WhatsApp messages
Books, confirms, reschedules, or cancels appointments
Provides accurate, consistent information
Routes complex or sensitive conversations to human staff
Maintains full conversation context and history

All interactions feel natural, professional, and brand-aligned.

Built for Real Businesses, Not Demos

Aventora is designed for production use in real organizations, including regulated industries.

Key characteristics:

Vendor-agnostic: No lock-in to AI models, telephony providers, or infrastructure
Deployment flexibility: Cloud, private cloud, or fully on-premise
Multi-language support: Serve diverse client bases effortlessly
Workflow-aware: Designed around real operational processes
Human-in-the-loop: AI knows when to escalate to staff

This makes Aventora suitable for organizations that require control, reliability, and long-term flexibility, not just experimentation.

Who Aventora Is For

Aventora Engagement Hub is used by organizations that rely on high-quality client interactions, including:

Legal practices
Insurance firms
Financial services
Real estate professionals
Medical and healthcare practices — see the public Healthcare segment guide

Each industry uses the same powerful core platform, with configurations tailored to their specific workflows and compliance needs. For industry packaging, see Supported Segments.

The Result

With Aventora, organizations achieve:

Faster response times
Fewer missed opportunities
Reduced administrative workload
Improved client experience
Better operational visibility
A scalable engagement model that grows with the business

Deployment Flexibility & Vendor Agnostic Architecture

Aventora Engagement Hub is designed with maximum flexibility and independence in mind. Unlike vendor-locked solutions, this platform can be deployed on any infrastructure that meets basic requirements, giving you complete control over your data, deployment environment, and technology choices.

Deployment Options

The platform can be deployed on any cloud provider (AWS, Azure, Google Cloud, IBM Cloud, Oracle Cloud, or any other), any data center of your choice, or even on-premises for maximum data sovereignty. This flexibility is essential for businesses with specific compliance requirements, data residency restrictions, or infrastructure preferences.

Whether you need to deploy in a specific geographic region for data sovereignty, meet regulatory requirements for on-premises deployment, or leverage existing infrastructure investments, the platform adapts to your needs rather than forcing you to adapt to vendor constraints. This deployment flexibility protects your investment and ensures you can evolve your infrastructure as your needs change.

LLM & Vendor Agnostic

The platform is completely LLM agnostic, meaning you can use any large language model provider you prefer. Support for multiple AI providers ensures you're never locked into a single vendor, allowing you to:

Switch AI providers when needed without changing your deployment
Use the best AI model for each specific use case
Leverage multiple AI providers simultaneously for redundancy or optimization
Choose providers that offer the best value and capabilities for your specific use case
Meet compliance requirements that may require specific AI providers or on-premises models

Similarly, the platform is vendor agnostic for telephony and other services. You can use any telephony provider that supports standard protocols, any database provider, any storage solution, and integrate with any external services through standard APIs. This vendor independence protects your investment and ensures you can evolve your technology stack as your needs change.

Infrastructure Independence

The platform's modular architecture means you're not dependent on proprietary cloud services or vendor-specific infrastructure. All components use standard protocols and can run on any infrastructure that supports common web technologies. This independence ensures:

No vendor lock-in - Your deployment remains portable and future-proof
Infrastructure optimization - Choose infrastructure that best meets your performance, security, and operational needs
Compliance flexibility - Deploy where you need to meet regulatory requirements
Scalability control - Scale infrastructure according to your needs and budget
Future-proof architecture - Adapt to new technologies without platform constraints

Core Features

The platform's core capabilities center around intelligent, natural conversations powered by advanced AI technology. These capabilities enable businesses to automate complex customer interactions that previously required human agents.

AI-Powered Phone Bot

At the heart of Aventora Engagement Hub is a sophisticated AI phone bot that enables natural conversations that feel authentic and human-like. The platform's vendor-agnostic architecture allows you to use any AI model provider you prefer, ensuring you can leverage the best available technology for your needs. This flexibility means customers can interact with your business using normal, conversational language rather than rigid menu systems or scripted responses.

The system delivers sub-second response times for most interactions, ensuring conversations flow naturally without awkward pauses that can make automated calls feel robotic. Context awareness means the AI remembers what was said earlier in the conversation, enabling natural multi-turn dialogues where follow-up questions are understood in context. This is essential for complex interactions like appointment scheduling, where customers might need to discuss dates, times, preferences, and special requirements in a fluid conversation.

Intent recognition capabilities enable the AI to understand what customers want even when they phrase requests in different ways. For example, whether a customer says "I'd like to schedule an appointment," "Can I book something?" or "I need to come in next week," the system recognizes they all have the same intent and responds appropriately. This natural language understanding makes interactions feel more human and reduces customer frustration.

Call Types

Different business needs require different types of automated calls. The platform supports four primary call types, each optimized for specific use cases and business objectives.

1. Appointment Booking

One of the most valuable use cases for AI phone automation is appointment booking, which can significantly reduce administrative workload while improving customer experience. The AI bot books appointments directly into your calendar, eliminating the need for manual entry and reducing the risk of scheduling errors.

The system handles date and time negotiation intelligently, understanding natural language like "next Tuesday afternoon" or "sometime next week" and converting them into specific calendar slots. When booking conflicts arise, the AI detects conflicts and resolves them by suggesting alternative times that work for both the customer and your schedule. After successful booking, the system sends calendar invites automatically, ensuring both parties have the appointment details in their calendars.

SMS and WhatsApp booking is optimized for text: the opening message greets the customer and asks what day works (no separate "is this a good time?" step). After the customer accepts an offered slot, the bot books immediately and sends one confirmation in the same thread — weekday, date, time, and timezone (for example, Friday, August 14 at 9:00 AM EDT). It does not send a second system SMS, include an internal booking ID, or promise another confirmation text. Voice bookings still use a second spoken confirmation, then a follow-up SMS with the same date/time format (no booking ID).

2. Confirmation Calls

No-shows are costly for businesses, wasting valuable appointment slots and revenue. Confirmation calls are a proven solution, with the platform helping businesses reduce no-show rates by up to 40% through automated reminders. The AI calls customers to confirm their upcoming appointments, and if they need to change their schedule, it can reschedule appointments on the spot.

The system sends reminder notifications via SMS or email as a backup, ensuring customers have multiple reminders of their upcoming appointments. This multi-channel approach significantly improves appointment attendance rates while freeing staff from manual reminder calls.

3. Informational Calls

Many customer inquiries are straightforward requests for information that don't require human interaction. Informational calls enable businesses to provide answers to frequently asked questions, share service details, or deliver important business information efficiently and cost-effectively.

These calls are perfect for sharing hours of operation, service descriptions, information about products or services, or any other standardized information that customers commonly request. By automating these routine interactions, businesses can handle more customer inquiries without increasing staff costs, improving efficiency while maintaining high-quality customer service.

Informational SMS replies can escalate to a conversation. When an informational SMS campaign asks an interest-seeking question and the customer replies with interest (and a team phone number is configured), the thread automatically upgrades to a conversational engagement on the same session — either notifying your team that the client is expecting a call, or asking whether they can talk now. Upgraded threads set notify_call_request so agreement to connect sends a call-request SMS to the Engagement callback number instead of placing a live conference bridge. The customer's first reply is always eligible to escalate; later replies escalate only while the thread is still fresh — within 24 hours of the campaign send. After that window, only a first reply upgrades. Pure notification messages (no question) never escalate, and opt-out requests are always honored.

4. Conversational Bridging

Sometimes customers need human assistance, but routing every call to a human agent defeats the purpose of automation. Conversational bridging provides the best of both worlds: the AI handles routine inquiries efficiently, but when a situation requires human judgment or expertise, it can connect customers with human agents seamlessly.

The handoff happens seamlessly through standard telephony conference capabilities, ensuring a seamless transition without dropped calls or customers having to call back. Most importantly, the system maintains conversation context during handoff, so human agents receive a summary of what the customer already discussed with the AI, eliminating frustrating repetition and ensuring continuity. Flexible call routing allows businesses to configure rules for when calls should be escalated to humans based on complexity, keywords, customer requests, or any other criteria that makes sense for your business.

SMS call-request notify (optional): For conversational SMS engagements you can set notify_call_request (CRM: Notify user by SMS instead of bridging). When the customer agrees to connect, Engagement Hub sends a secure call-request link SMS to your Engagement callback number (user_phone_number) — the same style of notification used for inbound callback requests — and completes the engagement with outcome call_requested instead of dialing a live conference. Informational SMS upgrades to conversational enable this mode automatically.

Multilingual Support

In today's global marketplace, supporting multiple languages is essential. The platform provides comprehensive multilingual support for 11 languages with native voice support, ensuring customers can communicate in their preferred language:

Supported languages include English, Spanish, French, German, Portuguese, Chinese, Arabic, Hindi, Japanese, Korean, and Farsi, with more languages being added regularly. The vendor-agnostic architecture supports multiple voice synthesis providers, allowing you to choose the best voice quality and provider for each language based on your preferences and requirements.

Each language uses high-quality, neural voice synthesis that sounds natural and professional, providing an authentic conversation experience regardless of language. The AI automatically detects the customer's language preference and responds accordingly, making multilingual support seamless and effortless for both businesses and customers.

Calendar Integration

Effective appointment scheduling requires seamless calendar integration. The platform provides comprehensive integration with the two most popular calendar systems, ensuring appointments are managed efficiently regardless of which calendar system your business uses.

Microsoft Outlook Integration

For businesses using Microsoft Outlook, the platform provides secure authentication that connects to Outlook calendars without requiring password sharing. This secure authentication method ensures calendar access is properly authorized while maintaining security. The integration supports multi-provider functionality, allowing businesses with multiple team members to manage appointments across all calendars from a single system.

Automatic token refresh ensures calendar connections remain active without manual intervention, and real-time calendar sync ensures the AI always has current availability information when scheduling appointments.

Google Calendar Integration

Similarly, Google Calendar integration uses secure authentication to connect Google calendars without password sharing. This authentication method ensures calendar access is properly authorized while maintaining security. The system supports multi-provider functionality for teams using Google Calendar, with automatic token refresh keeping connections active without manual intervention, and real-time calendar sync ensuring the system always has current availability information.

Advanced Calendar Features

Beyond basic integration, the platform includes sophisticated calendar management capabilities. Multi-provider availability checking enables the AI to find available time slots across multiple team members' calendars, perfect for businesses where appointments can be scheduled with any available provider.

Specialty filtering allows businesses to match customers with appropriate providers (e.g., general dentist vs. orthodontist), ensuring appointments are scheduled with the right person. Conflict detection prevents double-booking, and the system can automatically suggest alternative times when conflicts occur.

Automatic calendar invites are sent to both parties after booking, and flexible date handling allows the AI to understand natural language date references like "tomorrow," "next Monday," or "two weeks from now." Domain-wide availability checking enables businesses to see availability across all team members, and channel-specific availability allows different availability settings for phone calls versus text messaging.

Calendar-aware callback scheduling ensures callbacks are scheduled at times when team members are actually available, and availability conflict resolution provides intelligent alternative time suggestions when the requested time isn't available. Broker callbacks always use call working hours (not 24/7 messaging hours), even when the conversation is on SMS — so the bot will not confirm 7:00 PM if the call center closes at 6:15 PM. If a scheduled SMS callback still fires after hours, Hub texts that the center is closed and when you will call, instead of “connecting you now.”

Conversational and informational deferrals (queue for later, bulk ingest, broker-busy reschedule) use working hours + max concurrent calls only — no Google/Outlook calendar holds and no customer calendar invites until the engagement is appointment booking (or confirmational appointment flows). Provider calendar OAuth and use_calendar still apply to real appointment scheduling.

Call Management

Managing automated calls at scale requires sophisticated tools for initiation, tracking, and analysis. The platform provides comprehensive call management capabilities that give businesses full control and visibility over their automated communication efforts.

Call Initiation

Flexibility in how calls are initiated is essential for different use cases. The platform supports multiple call initiation methods:

Single call API allows programmatic initiation of individual calls, perfect for integrations or triggered calls based on specific events
Bulk call operations support CSV file uploads, enabling businesses to initiate hundreds or thousands of calls from a simple spreadsheet
Scheduled calls allow calls to be planned in advance, ensuring calls happen at optimal times for each customer
Callback requests enable customers to request calls at convenient times, improving engagement rates
Email Pull Service automatically monitors email inboxes and converts email inquiries into phone calls or messages, ensuring no customer inquiry goes unanswered regardless of how it arrives

Email Pull Service

For businesses that receive appointment requests or customer inquiries via email, the Email Pull Service automates the process of converting emails into phone calls or messages. Common use case: Meta Instant Form lead notifications — see Meta Ad Management.

The service uses automated email mailbox monitoring to continuously check email inboxes for new messages. AI-powered email content extraction intelligently extracts relevant information from emails, understanding natural language and varying email formats, ensuring that appointment requests and customer inquiries are accurately identified and processed even when formatted differently. Automatic call queuing converts extracted information into phone calls or SMS messages, ensuring email inquiries are handled promptly.

PDF attachment text extraction enables the system to read information from attached documents, and both account-level and user-level configurations provide flexibility for different organizational structures. Configurable pull intervals (30-3600 seconds) allow businesses to balance responsiveness with resource usage.

Email handling options (mark_read, delete, leave_unread) give businesses control over email processing, and processing logs and retry capabilities ensure reliability. Support for multiple call types from emails means different types of email inquiries can trigger different types of automated calls, and test number override enables development and testing without affecting production.

Domain-aware routing ensures emails are processed in the context of the correct business domain, essential for multi-tenant deployments.

Call Tracking

Comprehensive tracking is essential for understanding call performance and optimizing automation strategies. The platform tracks 15+ distinct call outcomes, providing detailed insights into call results. Every call includes full conversation transcription, enabling businesses to review exactly what was said and identify areas for improvement.

Duration tracking helps understand conversation length patterns, providing insights into how long different types of calls typically take and helping you optimize your automation strategy. Usage tracking provides visibility into how the system is being used, helping you understand patterns and optimize operations. Outcome analysis tools help identify trends, success patterns, and areas where the automation might need refinement, enabling continuous improvement of your phone automation strategy.

Call Outcomes

The platform tracks detailed call outcomes to provide actionable insights:

successful: Call completed successfully and achieved its objective
appointment_booked: An appointment was successfully booked
confirmed: An existing appointment was confirmed
rescheduled: An appointment was rescheduled to a different time
voicemail: Twilio AMD (or Hub detection) classified the call as an answering machine. Hub hangs up and does not leave a mailbox message. Informational phone calls may briefly show delivered until the AMD callback lands; the final persisted outcome is voicemail.
no_response: The call was not answered
call_failed: A technical failure occurred during the call
callback_requested: The customer requested a callback at a different time, including timed delays such as "do not call until after 11am"
call_requested: Conversational SMS with notify_call_request — staff was notified via secure-link SMS that the client is expecting a call (engagement completed)
user_declined: The customer declined the offer or request
cancelled: The call was cancelled before completion
do_not_call: The customer requested permanent opt-out (added to the do-not-call list). Timed delays ("don't call until 11am") are callback_requested, not DNC.
not_ready: The customer declined the offer but did not request permanent opt-out (may be contacted again later)
user_busy: The customer was busy and couldn't take the call
wrong_number: The number reached was incorrect
no_suitable_time: No suitable appointment time was found
incomplete: The call was started but not completed

This granular tracking enables businesses to analyze call performance in detail, identify patterns, and optimize their automation strategies.

Bulk Operations

For large-scale outreach campaigns, efficiency is critical. The platform's bulk operations capabilities enable businesses to process thousands of calls efficiently:

CSV file upload allows businesses to prepare call lists in spreadsheet format and upload them for processing. Batch processing handles large volumes efficiently, with progress tracking providing real-time visibility into campaign progress. Error handling ensures that individual call failures don't stop entire batches, and retry mechanisms automatically retry failed calls when appropriate.

Batch status monitoring provides comprehensive visibility into batch operations, including success rates, completion percentages, and detailed error reporting.

Dashboard & Analytics

Understanding call performance is essential for optimizing automation strategies and demonstrating ROI. The platform provides comprehensive dashboard and analytics capabilities that give businesses real-time visibility and detailed insights.

Real-Time Monitoring

The dashboard provides real-time monitoring capabilities that keep you informed of call activity as it happens. Active calls display shows which calls are currently in progress, while live call status updates provide instant notifications when call status changes.

Real-time balance tracking ensures you always have visibility into your account status, preventing unexpected interruptions. Call outcome notifications alert you immediately when important outcomes occur (like successful bookings), enabling timely follow-up actions. WebSocket updates ensure all information is current without requiring page refreshes, providing a seamless monitoring experience.

Analytics

Comprehensive analytics provide insights into call performance and business impact:

Total calls statistics show call volumes over time, helping identify trends and peak periods so you can optimize staffing and scheduling
Success rate metrics track the percentage of calls that achieve their objectives, enabling you to measure the effectiveness of your automation
Usage analytics provide visibility into how the system is being used, helping you optimize operations and resource allocation
Appointment booking stats show booking rates and help measure the return on investment of your phone automation efforts
Call duration averages help understand conversation length patterns, providing insights into customer engagement and conversation efficiency
Outcome distribution shows how calls are distributed across different outcomes, helping you identify areas for improvement

These analytics enable data-driven decision making, helping businesses understand what's working, identify areas for improvement, and demonstrate the value of automation to stakeholders.

Ask AI (campaign intelligence)

Path: Engagement Hub → Analysis (Call Analysis) → Ask AI

Ask AI is a feature-flagged campaign-intelligence workspace (ASK_AI_ENABLED). When enabled, date and campaign filters sit above two inner tabs:

Reports (default) — the 15 predefined report cards. Click the info icon on a card for the full explanation of what that report returns.
Conversation — a chat (same bubble/composer pattern as the site chatbot). Click the info icon next to the Conversation tab (or in the chat header) for question types and how to ask them. Ask a reporting question, then follow-ups like “break that down by campaign.” Prior turns are sent with each request; the thread is in-memory and clears when you leave Analysis or click New conversation. You can put a date range and channel in the question or use the filters and channel control. Follow-ups without a new date keep the previous question’s window (not a wider filter-bar range). Channel-specific counts use the first channel on the engagement. Questions about who received a Hub notification (staff recipient emails) and how those engagements ended (for example connected calls) join notification delivery to the engagement that fired the alert — the same join as Analysis → Reports → Notifications. Say both sides of that question (who was notified, and the engagement outcome); ask for recipient emails if you need the addresses as CSV. Answers are markdown text first; a preview table and CSV download attach to that turn only when you need a list. Signed-in admins may download contact-level CSV with fields they can already see in Call Details, and notification-recipient emails they can already see on the Notifications report. Phone numbers, emails, and transcripts are not sent to the AI model.

Public tokenized report links still use the predefined report catalog only (no Conversation tab, no custom contact-level export). Conversation reporting requires OPENAI_API_KEY.

LLM narrative summaries on canned report cards (when enough sample volume and OPENAI_API_KEY are available) use aggregates only—not raw transcripts. Some quality/objection reports use transcript classifications stored for your account.

Report	What it shows
Overall Engagement Summary	Top-level health: outreaches, reply / meaningful-reply / conversation / conversion rates, time to first reply, trend vs prior period, engagement funnel.
Campaign / Email Config Comparison	Side-by-side ranking of selected campaigns on reply quality, speed, and negative rates.
Responder vs Non-Responder Pattern Analysis	How people who reply differ from those who don’t (hours, channels, attempt counts).
Best Day and Time to Generate Replies	Reply rates by hour and weekday in the account timezone; business vs after-hours.
Retry Effectiveness Analysis	Reply lift (and negatives) by attempt number; suggested max attempts.
Retry Delay Optimization	Whether wait time between attempts helps or looks aggressive.
Channel Effectiveness Analysis	First-touch SMS / phone / email performance.
Channel Sequence Analysis	Multi-channel path ordering (e.g. SMS→phone) by meaningful-reply rate.
Conversation Quality Analysis	Meaningful vs low-quality / negative reply types from classifications.
Objection and Reason Analysis	Top objections after a reply (pricing, timing, already insured, etc.).
First Message Effectiveness Analysis	Which opening-message patterns earn meaningful replies.
Conversation Momentum Analysis	Speed to first reply and how deep threads go before they stall.
Contact Fatigue / Spam Risk Analysis	Over-touch and negative-after-retry risk score with recommended caps.
Campaign Interference / Overlap Analysis	Same contacts hit by multiple campaigns in 3–30 day windows.
Single Biggest Improvement Recommendation	One highest-impact next change synthesized from comparison, timing, retries, and fatigue.

Public tokenized report links (email-import customers) use the same report catalog when Ask AI access tokens are provisioned.

Reports

Beyond real-time monitoring, Call Analysis → Reports provides operational and conversation reporting as tabs on that panel:

Operational — call statistics, details, optional status history and transcripts; export as PDF or Excel
Conversations — opens the full report in the same window with colored summary metrics, reply-rate and outcome charts, and conversation cards with customer details and chat-style transcripts (customer-reply engagements only). PDF export from the report page. Summary metrics for the selected period include total engagements, active (with replies), negative responses (do not call / not ready), positive responses (connected / callback requested), timed out (no reply after max retries or informational with no reply), and others (in progress, etc.)
Lead Score — scored conversational leads with band and reasoning; export to Excel
Notifications — per-recipient delivery for Engagement Hub alert rules and the engagement outcome that triggered each alert; Excel export
Epic Producer — upload an Applied Epic Calls-Producer CSV/Excel (Queue Name, Producer, Team, Phone number IN, Clients number, Phone type, Start Time, Talk Time). Engagements in the selected date range are matched to producer calls by normalized client phone. The workbook keeps those Epic columns, adds Aventora name/channel/outcome/timestamps, a matched-only producer×queue pivot, a summary, and unmatched Epic calls. Rows where Clients number equals the inbound DID are skipped. An attribution checkbox (on by default) only counts Epic calls on or after the Aventora engagement. This report attributes producer call activity, not closed policy premium.

Custom date ranges and channel filters (single, batch, email pull) apply across these reports. These capabilities make it easy to create presentations for stakeholders and track progress toward business objectives.

Account Management

Enterprise deployments require sophisticated account management capabilities that support teams, multi-tenant architectures, and complex organizational structures.

Account Features

The platform supports multi-account functionality, enabling businesses to organize their operations across multiple accounts when needed. Team member management allows you to add team members with appropriate permissions, while role-based permissions ensure team members only have access to features they need.

Account settings provide centralized configuration management, enabling you to configure all aspects of your phone automation system from a single location. Usage visibility provides transparency into system usage patterns, helping you understand how the platform is being utilized. Domain-to-account mapping enables integration with domain-based systems, and external user ID linking allows connections to external user management systems.

Domain Accounts & Multi-Tenant Support

For businesses managing multiple domains, brands, or customer segments, the platform provides sophisticated multi-tenant support:

Domain-based account resolution automatically routes calls and configurations based on domain context. Multi-tenant architecture ensures complete data isolation between domains while sharing infrastructure efficiently. Domain-specific settings and configurations allow each domain to have customized behavior, and domain-level telephony credentials enable separate phone numbers and configurations per domain, ensuring each business can maintain its own telephony infrastructure.

Domain-wide calendar availability checking allows businesses to see availability across all team members within a domain, and specialty-based provider filtering ensures customers are matched with appropriate providers. Domain admin account lookup simplifies account management, and both account-level and user-level configurations provide flexibility for different organizational structures.

Provider Management

For businesses with multiple service providers (like medical practices with multiple doctors), provider management is essential:

Add and remove providers easily as your team changes
Specialty configuration ensures customers are matched with the right provider
Enable and disable providers temporarily (e.g., for vacations) without removing them
Calendar linking connects each provider's calendar to the system
Provider-specific settings allow customization per provider
Multi-calendar support in domain mode enables comprehensive availability checking
Availability type configuration allows different availability for calls versus messaging

Usage Tracking & Management

Understanding how your phone automation system is being used is essential for optimization and business planning. The platform provides comprehensive usage tracking and management capabilities that give you full visibility into your operations.

Usage Tracking

The platform provides detailed usage tracking that shows how the system is being used, including call volumes, call types, channel usage (phone, SMS, WhatsApp), and outcome patterns. This visibility enables you to understand your customer communication patterns and optimize your automation strategy accordingly.

Real-time balance tracking ensures you always have visibility into your account status, preventing unexpected interruptions. Usage analytics provide insights into consumption patterns, helping you optimize resource allocation and identify trends.

Usage Management

Purchase history provides a complete record of all transactions, enabling you to track spending over time and plan for future needs. Usage reporting helps you understand how the platform is being consumed, making it easy to budget and allocate resources effectively.

The platform's flexible architecture means you can optimize usage based on your business model and preferences, whether you prefer pay-per-use models, subscription-based approaches, or custom pricing arrangements with your deployment provider.

API Integration

For businesses that need to integrate phone automation into existing systems, comprehensive API capabilities are essential.

RESTful API

The platform provides a comprehensive RESTful API with 50+ endpoints covering all functionality, from call initiation to analytics retrieval. Comprehensive documentation includes detailed endpoint descriptions, request/response examples, and authentication instructions.

SDK examples in multiple programming languages help developers get started quickly, and webhook support enables real-time notifications for events like call completions or appointment bookings. Rate limiting protects API stability and ensures fair usage across all customers.

Authentication

Secure API access is provided through API key authentication that's simple to implement, making it easy for developers to integrate the platform into existing systems. Enterprise authentication support provides additional security options for organizations with strict authentication requirements. Scoped permissions allow API keys to be limited to specific operations, following security best practices by ensuring integrations only have access to the functionality they need. Token management provides secure, time-limited access, automatically expiring tokens after set periods to minimize security risks.

API Key Management

Comprehensive API key management enables secure integration:

Create, read, update, and delete API keys through both the dashboard and API
Key expiration management allows keys to be set with expiration dates for improved security
Permission-based access control limits what each key can do
Key revocation and permanent deletion enable immediate response to security incidents
Automatic cleanup of expired keys maintains security hygiene
Key metadata including name, description, and permissions helps organize multiple keys
Created-by tracking provides audit trails
Active/inactive status management enables temporary key deactivation

Webhooks

Real-time event notifications are essential for integrations. The platform supports webhooks for key events:

Call status updates notify you when calls start, complete, or change status
Appointment bookings trigger notifications when appointments are successfully booked
System alerts notify you of important events and status changes
Custom event notifications can be configured for specific business needs

SMS Integration

Text messaging has become a preferred communication channel for many customers. The platform provides comprehensive SMS capabilities that complement voice calling.

The system can send SMS messages to customers, enabling communication through their preferred channel. Two-way SMS conversations allow customers to respond and engage in natural text-based dialogues with the AI. SMS notifications can be sent for appointments, reminders, or important information, and bulk SMS support enables large-scale text campaigns.

Delivery tracking ensures you know when messages are delivered, and session management with a 30-minute timeout maintains conversation context across multiple messages. Conversation history tracking provides a complete record of SMS interactions, and automatic retry scheduling for voicemails or no-response scenarios ensures important messages are eventually delivered.

Pre-send phone validation (Twilio SMS)

When enabled, outbound Twilio SMS runs Twilio Lookup v2 (line_type_intelligence) before the first send to each destination number. Invalid numbers, landlines, and fixed VoIP lines are blocked with a failed engagement outcome instead of attempting delivery.

Server default: Off unless your platform operator enables validation at the server level.
Per-account override: Settings → Phone → Account Settings → Calls → Validate Phone Numbers Before SMS. When set, the account toggle overrides the server default for that tenant.
Cost control: Lookup results are cached by phone number; each destination is validated at most once across engagements.
Scope: Twilio SMS only (not Telnyx, device relay, or customer-reply messages in an active SMS thread).

Cancelling in-progress SMS conversations

When you cancel an in-progress conversational SMS from Analysis → Details (or the dashboard), outbound processing stops: the row is marked Cancelled, queued retries are skipped, and a Cancelled event is recorded in Conversation History.

If the customer replies after cancel, the engagement reactivates automatically: the AI (or agent) can respond, status returns to In Progress (or Queued when a callback is scheduled), and the history timeline shows Cancelled then Resumed. This matches live SMS session behavior—the cancel flag blocks outbound queue work but does not discard an active customer thread.

Repairing rows stuck before this fix: run scripts/repair_cancelled_sms_post_customer_reply.py (dry-run by default) from the Engagement Hub deployment package, then --execute after review. Use --call-sids to target specific conversations.

WhatsApp Integration

With over 2 billion users worldwide, WhatsApp has become a critical communication channel for businesses. The platform provides comprehensive WhatsApp support through standard WhatsApp Business API integrations.

WhatsApp Messaging

WhatsApp text messages enable businesses to reach customers on the world's most popular messaging platform. Two-way WhatsApp conversations allow customers to respond and engage naturally, and template message support enables pre-approved messages for notifications and confirmations, ensuring compliance with messaging platform requirements.

Session management reuses the SMS session infrastructure, ensuring consistent behavior across messaging channels and simplifying operations. Conversation history tracking maintains complete records of WhatsApp interactions, enabling you to review conversations and improve your automation over time. Domain-specific WhatsApp number configuration allows each business domain to have its own WhatsApp number, ensuring proper branding and customer recognition.

WhatsApp Voice Calls

Beyond messaging, the platform supports outbound WhatsApp voice calls, bringing AI voice capabilities to WhatsApp users. Full bot conversation support means all call types (booking, confirmation, informational, conversational) work through WhatsApp voice, and standard telephony protocols ensure reliable call handling.

Domain-specific configuration allows customized behavior per domain, and display number customization enables businesses to control how their number appears to customers.

Multi-Channel Support

Modern businesses need to meet customers on their preferred communication channels. The platform provides unified multi-channel support that maintains consistent functionality across channels while optimizing for each channel's unique characteristics.

The platform supports four primary channels:

Phone: Traditional voice calls provide the most personal interaction experience
SMS: Text messaging offers convenience and accessibility
WhatsApp Message (wmsg): WhatsApp text messaging reaches international customers effectively
WhatsApp Voice (wvoice): WhatsApp voice calls bring voice AI to WhatsApp users

All channels support the same call types and features (appointment booking, confirmation, informational, conversational), ensuring consistent functionality regardless of how customers choose to communicate. However, each channel includes channel-specific optimizations that take advantage of each channel's unique capabilities and user expectations.

Security & Privacy

Security and privacy are foundational principles of Aventora Engagement Hub. As a system that handles sensitive customer conversations, appointment information, and potentially regulated data, comprehensive security and privacy protections are built into every layer of the platform. The platform is designed to meet enterprise security requirements and support compliance efforts for frameworks such as GDPR and HIPAA-ready healthcare deployments (formal compliance requires appropriate agreements and configuration).

The platform's vendor-agnostic and infrastructure-flexible design enhances security by allowing you to deploy in environments that meet your specific security requirements, whether that's a private cloud, a specific geographic region, or on-premises infrastructure. This control over deployment location is critical for businesses with data sovereignty requirements or strict compliance needs.

AI and Customer Data

Aventora uses third-party AI model providers (such as OpenAI and Groq) to power conversations. Customer conversation content is sent to configured providers for inference during active sessions. Provider data handling, retention, and training practices depend on your chosen provider and contract terms—review provider documentation and contact sales@aventora.ai for enterprise deployment options (including provider choice, region, and on-premises models where supported). See Security Overview and Integration Security for data-flow details.

Data Security

The platform is designed with HIPAA-ready architecture in mind, providing security controls and audit capabilities that support healthcare use cases when deployed with appropriate Business Associate Agreements and configuration. Enterprise-grade encryption protects data in transit; encryption at rest depends on your deployment environment. Secure credential storage ensures sensitive information like API keys and calendar tokens are protected using industry-standard practices.

Audit logging captures all system events for security auditing and compliance, providing complete visibility into system usage and access patterns. Data isolation ensures that different customers' data cannot be accessed across account boundaries, maintaining privacy and security in multi-tenant deployments.

Access Control

Comprehensive access control ensures only authorized users can access system capabilities:

API key management provides secure programmatic access with fine-grained permissions
Role-based permissions ensure users only have access to features they need, following the principle of least privilege
Account isolation prevents cross-account data access, maintaining data privacy and security
Session management provides secure user sessions with proper authentication and authorization
Token expiration limits the window of vulnerability if tokens are compromised, automatically invalidating tokens after set time periods

Privacy & Data Protection

Privacy regulations like GDPR require businesses to carefully manage how they collect, store, and use customer data. The platform includes features designed to help businesses meet these requirements:

Data minimization principles ensure only necessary information is collected
User consent management capabilities enable proper consent tracking
Privacy controls allow users to manage their privacy settings
Data subject rights support enables fulfilling requests for data access, deletion, and portability
Data retention policies can be configured to automatically manage data lifecycle
Data deletion capabilities enable permanent removal of data when required

Advanced Features

Beyond core functionality, the platform includes advanced features that optimize performance, improve user experience, and enable sophisticated use cases.

Voicemail Detection

When calls go to voicemail, the system needs to respond appropriately. Automatic voicemail detection identifies when a call has reached voicemail, enabling the system to leave appropriate messages. Custom voicemail messages can be configured per call type or domain, and message recording ensures voicemails are captured for review.

Delivery confirmation verifies that voicemail messages were successfully left, and configurable retry intervals allow businesses to control when follow-up attempts are made. The Phone2SMS fallback option can automatically send an SMS when voicemail is detected, ensuring important messages are delivered.

Phone2SMS Feature

Sometimes customers prefer text messaging over phone calls. The Phone2SMS feature provides automatic fallback from phone to SMS when voicemail is detected, ensuring messages are delivered through the customer's preferred channel.

The feature uses account-level preference settings to control behavior, and provides immediate SMS callback when voicemail is detected, creating a seamless experience. Seamless channel switching maintains conversation context across channels, and domain-aware preference resolution ensures settings are applied correctly in multi-tenant deployments.

Enhanced Retry & Callback Logic

Not all calls succeed on the first attempt. Sophisticated retry logic ensures important calls eventually reach customers:

Configurable retry intervals at the domain level allow businesses to control retry timing based on their needs and customer preferences. Incomplete call retry scheduling automatically retries calls that didn't complete, while voicemail retry scheduling retries calls that went to voicemail.

Maximum retry attempts tracking prevents excessive retry attempts, and automatic callback scheduling enables customers to request callbacks at convenient times. Duplicate callback prevention ensures customers aren't called multiple times for the same request, and conversation chain tracking using original_call_sid maintains context across retries.

The system supports configurable retry parameters:

max_retry_attempts: Maximum number of retry attempts per conversation (default: 4 attempts)

Live Call Bridging

Sometimes customers need human assistance during an AI conversation. Live call bridging provides instant human handoff without requiring the customer to hang up and call back.

The feature uses standard telephony conference capabilities to seamlessly add a human agent to the call, creating a seamless transition that feels natural to the customer. Most importantly, context preservation ensures the human agent receives a summary of what the customer already discussed with the AI, eliminating frustrating repetition and ensuring continuity.

Email Integration

Email remains an important communication channel for businesses. The platform provides comprehensive email integration capabilities:

Email notifications can be sent for appointments, confirmations, and other events. Calendar invite delivery ensures customers receive calendar invitations via email, and appointment confirmations can be sent as backup to phone confirmations. Custom email templates allow businesses to brand emails appropriately.

Analytics & Reporting

Understanding performance and demonstrating ROI requires comprehensive analytics:

Comprehensive analytics provide insights into call performance, success rates, and usage patterns
Custom reports enable analysis tailored to specific business needs
Export capabilities allow further analysis in external tools
Performance metrics track system performance and reliability
Usage trends help identify patterns and optimize strategies

Greeting Service

First impressions matter in customer interactions. The Greeting Service provides AI-generated greetings that set the right tone for each conversation.

The service generates greetings for all call types, ensuring each interaction starts appropriately. Language-specific greeting generation ensures greetings are natural in the customer's language, and context-aware introductions customize greetings based on call purpose and customer information.

Instruction-based customization allows businesses to specify how greetings should sound, and direct instruction support (pass-through mode) enables complete control when needed. Channel-optimized greetings ensure greetings work well for phone calls versus SMS/WhatsApp messaging, and provider name integration personalizes greetings for multi-provider businesses.

Appointment details integration for confirmational calls ensures greetings mention the specific appointment being confirmed, making conversations feel more personal and professional.

Thinking Ambient Audio

Even with sub-second response times, AI processing can sometimes take a moment. A continuous ambient audio loop — a soft harmonic pad with a gentle breathing pulse — plays while the LLM is working, eliminating awkward silence that can make automated calls feel broken.

The audio is generated programmatically — a warm, musical texture using layered harmonics — so it requires no TTS calls, no pre-generation warmup, and works for every language and voice. It starts 2 seconds after the user stops speaking and stops immediately when the user speaks again or the LLM responds, ensuring a seamless conversational flow.

Immediate audio feedback during processing ensures customers know the system is working, and reduced perceived latency makes conversations feel faster even when processing takes a moment.

Information Retrieval Adapter

For informational calls, the system needs to retrieve relevant information about customers or topics. The Information Retrieval Adapter provides a pluggable system for retrieving this information.

The adapter system uses a default LLM-based mock implementation for testing and simple use cases, but supports customer information context and request-based information retrieval for production deployments. Custom adapter support enables integration with existing customer databases, CRM systems, or other information sources.

This capability is particularly useful for informational calls from email pull, where the system needs to retrieve specific customer information to provide personalized responses.

Technical Features

Beyond feature capabilities, the platform includes technical features that ensure performance, reliability, and scalability.

Performance

Performance is critical for user experience, especially in real-time voice conversations:

Sub-second response times for most interactions ensure natural conversation flow
Connection pooling (1-20 connections) optimizes database performance
Service caching reduces redundant operations and improves speed
Optimized database queries ensure fast data access even with large datasets
Horizontal scaling support enables performance scaling as usage grows
Thinking chime fills silence during LLM processing with zero TTS overhead
In-memory caching for hot data provides instant access to frequently used information
Pre-generated greetings eliminate generation delays
Common phrases caching speeds up frequent responses

Reliability

Business-critical applications require high reliability:

99.9% uptime target ensures service availability
Automatic failover handles failures gracefully without service interruption
Health monitoring detects issues before they impact users
Error recovery automatically recovers from transient errors
Backup systems provide redundancy for critical components

Scalability

As businesses grow, the platform must scale with them:

Unlimited concurrent calls means the system can handle any call volume
Horizontal scaling allows adding capacity by adding servers
Load balancing support distributes load efficiently across servers
Database optimization ensures performance at scale
Resource management prevents resource exhaustion even under heavy load

Monitoring

Visibility into system health and performance is essential:

Health check endpoints enable automated monitoring
Logging system provides comprehensive operational logs
Error tracking helps identify and fix issues quickly
Performance metrics track system performance over time
Alert system notifies administrators of critical issues

Integration Features

The platform is designed to work with your existing technology infrastructure.

Third-Party Integrations

The platform integrates with leading third-party services through standard APIs, providing flexibility in vendor selection:

Telephony providers enable voice calling, SMS, and WhatsApp capabilities
AI model providers power the conversation capabilities
Speech-to-text services provide voice recognition
Text-to-speech services provide natural voice synthesis
Microsoft Graph API enables Outlook calendar integration
Google Calendar API enables Google Calendar integration
Payment processors can be integrated based on your preferences and requirements
Domain-Chatbot API provides multi-tenant support and integration
IMAP enables email pull service functionality

Custom Integrations

For businesses with unique requirements, the platform supports custom integrations:

Webhook support enables real-time event notifications to external systems
RESTful API provides programmatic access to all functionality
SDK availability in multiple languages simplifies integration development
Custom endpoints can be developed for specific use cases
Plugin architecture enables extending functionality when needed

Compliance Features

Enterprise deployments often require compliance with various regulations and standards.

HIPAA-Ready Architecture

For healthcare use cases, the platform provides HIPAA-ready architecture:

Healthcare data protection through encryption and access controls (when appropriately deployed)
Secure data transmission via encrypted connections
Access controls to limit data access to authorized users
Audit trails for compliance review
Business Associate Agreements can be established where required for regulated healthcare deployments

Data Privacy

Privacy regulations like GDPR require careful data management:

GDPR-aligned features to help businesses meet European privacy requirements
Data encryption protects personal information
Privacy controls allow users to manage their data
Data retention policies can be configured to meet requirements
User data rights support enables fulfilling data access and deletion requests

Developer Features

For businesses that need to integrate or customize the platform, comprehensive developer tools are available.

API Documentation

OpenAPI/Swagger documentation provides comprehensive API reference
Code examples in multiple languages help developers get started
SDK libraries simplify integration in common programming languages
Integration guides provide step-by-step instructions for common scenarios
Testing tools help validate integrations

Development Tools

Local development setup enables testing without production deployment
Docker support simplifies development environment setup
Testing framework helps ensure code quality
Debugging tools assist in troubleshooting issues
Logging system provides visibility during development

Domain-Level Settings

The platform provides extensive configuration options that allow businesses to customize behavior for their specific needs.

Configurable Settings

Domain-level settings provide fine-grained control over phone automation behavior, allowing you to customize how the system operates for your specific business needs:

Default call duration: Set the expected length of calls in minutes, helping the system optimize scheduling and resource allocation
Voicemail retry interval: Configure how many hours the system waits before retrying calls that went to voicemail, allowing you to balance persistence with customer preferences
Incomplete call retry interval: Set the wait time before retrying calls that didn't complete, ensuring important messages eventually reach customers
Maximum retry attempts: Control the maximum number of retry attempts per conversation, preventing excessive calling while ensuring important messages are delivered
Comfort audio at call start: Enable or disable audio chimes or greetings at the beginning of calls, providing customers with immediate feedback that the call is active
WhatsApp number: Configure the WhatsApp Business number used for your domain, ensuring customers recognize your business number
Display number: Set the phone number that appears to customers for outbound calls, helping maintain brand consistency and customer recognition

Settings follow a priority hierarchy that provides flexibility for different deployment scenarios. Domain-level settings override account-level settings, which in turn override default system settings. This hierarchy enables businesses to customize behavior at the most appropriate level for their organization structure.

Domain-Specific Features

Multi-tenant deployments benefit from domain-specific features:

Domain-level telephony credentials (per channel) enable separate phone numbers and configurations
Domain-wide calendar availability provides visibility across all providers in a domain
Domain-specific billing allows separate billing per domain
Domain routing for calls ensures calls are processed in the correct context
Specialty-based filtering enables provider matching within domains

Administration Dashboard

The Aventora Admin Dashboard provides comprehensive management capabilities for all platform features, enabling non-technical users to configure, monitor, and optimize their phone automation deployments.

Phone Assistant Administration

The dashboard provides extensive administration capabilities for phone automation features.

Account Settings Management

Calendar integration settings for Google and Outlook OAuth provide straightforward setup and management of calendar connections. The same OAuth connection can be created in Engagement Hub admin or Aventora CRM (Settings → Accounts); tokens are stored once in Hub and reused across calendar booking, inbound email (when configured with a connected account), and CRM calendar/email sync. Users who connected in the other product see an Enable sync option instead of signing in again.

Answer domain questions during conversations — under Settings → Phone → Account Settings → Calls, enable Answer Domain Questions During Conversations (conversational_domain_knowledge_enabled). When on, conversational engagements on phone, SMS, WhatsApp, and voice can answer factual customer questions from your Domain Assistant knowledge base (domain-chatbot POST /query/agent) while the bot continues working toward connecting the customer with your team. Availability replies (“yes”, callback times), opt-out requests, and bridge actions always follow the normal engagement flow. Requires a provisioned account domain_chatbot_api_key and an indexed knowledge base. Optional env: CONVERSATIONAL_DOMAIN_KNOWLEDGE_WAIT_SECONDS (default 3), CONVERSATIONAL_DOMAIN_KNOWLEDGE_QUERY_TIMEOUT_SECONDS (default 8).

Changelog

Date	Change
2026-08-21	Outbound AMD outcome: When Twilio AMD reports a machine (`machine_start` and other machine/fax classifications), the final `call_logs` outcome is `voicemail`, even if informational `/start` first stored `delivered`. Hub still hangs up and does not leave a mailbox message. Duplicate AMD callbacks stay `voicemail`; a later `delivered` write cannot overwrite it.
2026-08-19	Ask AI Conversation help: info icon opens question types and asking guidance, including joining Hub notification recipients to engagement outcomes.
2026-08-19	Ask AI tabs: Reports is the first/default tab (15 canned cards); Conversation is second. Follow-ups keep the previous question’s date range; JSON envelopes are not shown in chat.
2026-08-19	Ask AI Conversation tab: Free-form reporting is a chat with follow-ups (in-memory). The 15 canned reports live on a Reports tab. Channel-specific counts use first channel.
2026-08-18	Call Analysis → Reports → Epic Producer: upload an Applied Epic Calls-Producer export and download a consolidated Excel that matches Aventora engagements to producer calls by client phone (DID-equals-client rows skipped; optional “Epic after engagement” attribution).
2026-08-13	SMS appointment booking: SMS/WhatsApp bookings confirm once in the same thread using weekday, date, time, and timezone. No second system SMS, no internal booking ID, no extra “To confirm?” after the customer accepts a slot, and the opener asks what day works instead of “is this a good time to chat?”. Voice bookings still use a spoken re-confirm plus a follow-up SMS in the same date format.
2026-08-13	Aventora Calendar empty hours: Domain Assistant persists Monday–Friday 9:00 AM–5:00 PM when a user has no availability records, and Hub booking offers the first remaining slot after the current local time.
2026-08-12	Call Analysis → Reports: reports are tabs (Operational, Conversations, Lead Score, Notifications) instead of stacked accordion rows.
2026-08-12	Broker callbacks use call hours: SMS/WhatsApp callback times are validated against call working hours (not 24/7 messaging hours). After-hours scheduled SMS callbacks notify the customer of the next opening instead of texting “connecting you now” when no broker can take the call.
2026-08-12	Timed delay vs DNC: Replies such as "do not call until after 11am" are classified as `callback_requested` (reschedule), not `do_not_call`. Permanent opt-out still applies to STOP, list-removal, and bare "do not call" / "do not call me again".
2026-08-11	Ask AI campaign intelligence: Documented all 15 predefined Call Analysis → Ask AI reports (what each shows). UI cards include an info icon with the full description; feature remains gated by `ASK_AI_ENABLED`.
2026-07-31	Google OAuth scopes: Unified Google connect (Hub and CRM) requests only verified consent-screen scopes — identity, `profile.emails.read`, `gmail.readonly`, and `calendar.events.owned.readonly` (no `gmail.send` / full calendar write). Appointment booking that creates Google Calendar events requires a broader verified scope set if re-enabled later.
2026-07-22	Supported Segments: Public Healthcare segment guide at Healthcare (patient engagement beside EMR; not the NDA proposal PDF).
2026-07-19	Email OOO pause (SES signature match): Amazon SES rewrites Gmail OOOs to `MAILER-DAEMON` with no In-Reply-To. Hub now matches the vacation signature (name tokens / phone) against round-robin email destinations, and SysInbound also reads `INBOX.Trash` where hosts often file these auto-replies.
2026-07-19	Email OOO pause after prior redispatch: A previous OOO redispatch for the same notification subject no longer blocks pausing a new unre-dispatched send. Attribution requires the OOO From (or embedded From in SES wrappers) so duplicate SES notices cannot pause the next round-robin recipient.
2026-07-19	Notification delivery failure + PAUSE/RESUME: Email bounces and failed SMS (wrong/undeliverable) pause that recipient and redispatch like OOO. Staff SMS keywords are `PAUSE` / `RESUME` (legacy `OUT` / `BACK` still accepted). Carrier `STOP` remains permanent opt-out.
2026-07-18	Notification OOO redispatch: When an email OOO is detected, Hub pauses that recipient and resends the same notification to the next active round-robin address.
2026-07-18	Notification round-robin pause (SysInbound): When Account Settings → Email uses System, OOO pause reads replies via domain-chatbot SysInbound IMAP (`SYS_INBOUND_*`), with Reply-To set on system outbound.
2026-07-18	Notification round-robin pause: SMS recipients can reply `PAUSE` / `RESUME` to leave or rejoin the rotation. Email out-of-office or bounce auto-replies pause that address (default 7-day TTL) when Hub can read the send mailbox (Google/Microsoft linked inbox, or System + SysInbound). Failed SMS delivery pauses and redispatches similarly. Admins can view and Resume paused recipients under Engagement Hub → Notifications.
2026-07-16	Conversations Report PDF: Export now matches the on-screen report styling — colored metric cards, reply-rate and outcome charts, and chat-style conversation cards with transcripts (not plain tables).
2026-07-16	Conversations Report: Under Call Analysis → Reports, renamed Operational Reports (plural) and added Conversations Report — opens in a dedicated page with colored metrics, charts, conversation cards, and PDF export. Summary includes period, total / active / negative / positive / timed out / others; detail lists only conversations where the customer replied.
2026-07-14	SMS call-request notify: Conversational SMS can set `notify_call_request` so customer agreement to connect sends a secure call-request SMS to `user_phone_number` instead of a live conference; outcome is `call_requested`. Informational→conversational SMS upgrades enable this automatically. CRM Engage dialog and campaign AUTO_ENGAGE expose the toggle for conversational SMS.
2026-07-13	Informational SMS escalation window: Later customer replies (2nd turn onward) to an interest-seeking informational SMS campaign can now upgrade to a conversational bridge while the thread is still fresh — within 24 hours of the campaign send. The first reply remains eligible at any time; after 24 hours only a first reply upgrades. Requires a configured team phone number; opt-out and notification-only messages are unaffected.
2026-07-08	Conversational domain knowledge: New account setting Answer Domain Questions During Conversations under Account Settings → Calls. When enabled, factual questions during conversational engagements are answered from Domain Assistant RAG in parallel with the bridge/callback flow; engagement actions always take priority.
2026-07-06	SMS cancel repair script: `scripts/repair_cancelled_sms_post_customer_reply.py` backfills rows cancelled before the customer-reply reactivation fix (dry-run by default; `--execute` to apply). See `scripts/find_cancelled_sms_post_customer_reply.sql` for preview SQL.
2026-07-06	SMS cancel + customer reply: Cancelling an in-progress SMS stops outbound/queue processing and records a Cancelled history event. If the customer replies afterward, the engagement reactivates (Resumed in history), status updates correctly, and callbacks can be scheduled again.
2026-07-05	Linked Email Pull to Meta Ad Management for Meta Instant Form lead ads.
2026-07-05	Conversational/informational scheduling: Deferred and queued conversational and informational engagements no longer create calendar events or send customer invites. Availability uses working hours (call or messaging) and max concurrent calls; calendar integration remains for appointment booking and confirmational flows.
2026-07-04	Async bulk ingest: CSV upload and JSON submit return 202 immediately; rows are written to the outbound queue in background (`bulk_ingest_jobs`). Batch status `ingesting` until complete, then `processing`.
2026-07-04	Bulk outbound pacing: Bulk worker and timeout retries now run parallel `/start` calls capped by `BULK_CALL_MAX_CONCURRENT` (default 20) instead of serial 5/min pacing. Legacy `BULK_CALL_RATE_LIMIT` optional when > 0.
2026-07-04	SMS/campaign retry scheduling: The timeout worker now scans oldest active (`In Progress`) conversations first and skips cancelled or paused rows when choosing which sessions to evaluate for retry. Fixes bulk campaigns where thousands of simultaneous active SMS sessions caused eligible retries to be starved behind newer rows.
2026-07-03	Caller ID verification in admin: Verify with Twilio on Settings → Profile and Account Settings (Default Broker Phone, Display Number). Registers numbers as Twilio Verified Caller IDs via voice call + validation code; verification is per E.164 on the tenant Twilio account.
2026-07-03	DNC detection (SMS/phone): Messages asking to be removed from a contact list or to stop all calls/texts/emails are classified as `do_not_call` (added to the DNC list), not `not_ready`. Includes new `handle_do_not_call` handler and expanded opt-out phrase matching.
2026-07-02	Domain validation (bulk): `/start` caches successful domain checks for 5 minutes and skips redundant domain-chatbot users HTTP when metadata succeeds with the account API key.
2026-07-02	Twilio SMS rate limit default — account-level cap is 1000 SMS/hour when not overridden in account settings (was 300). Bulk uploads still do not count until messages are sent.
2026-07-02	Twilio SMS/call rate limits — bulk CSV uploads no longer consume hourly SMS limits before messages are sent; limits apply to actual outbound attempts only.
2026-07-02	Call availability toggle — Available / Away on mobile Home and Admin Account Settings → Calendar → Call Availability; overrides phone routing without editing weekly hours.
2026-06-30	SMS phone validation: optional Twilio Lookup before outbound Twilio SMS; per-account toggle under Settings → Phone → Account Settings → Calls → Validate Phone Numbers Before SMS. Blocks invalid, landline, and fixed VoIP numbers; caches one lookup per phone number.

Executive Overview​