Monobot

What is Monobot?

Monobot is an enterprise AI voice and chat agent platform provided by TRING, Inc., designed for use in customer service, sales, scheduling, and contact center automation.

Users can configure proxies, knowledge bases, voices, workflows, and external integrations through a visual interface, and then deploy the proxies to web chats, web calls, or phone numbers.

Classifying it as an AI programming tool based on the old information in the database is inaccurate. Monobot’s core is a conversational customer experience platform, and its development capabilities are primarily aimed at integration and process expansion.

Main functions

No-code AI agent creation

• Templates and custom proxies: You can start with templates for sectors such as transportation, healthcare, banking, real estate, retail, and education, or you can begin with blank proxy configurations.
• Tips and introductory remarks: The administrator sets the proxy name, welcome message, roles, business objectives, boundaries for responses, and processing steps, after which testing can be carried out directly in the management dashboard.
• Shared process for voice and chat: The same agent can handle text chats, web-based voice conversations, and phone calls, but the input requirements, latency, and costs vary for each channel.
• Multilingual support and voices: The platform supports over 20 languages and various voices; it is possible to choose between TTS and STT processors, as well as to adjust the pronunciation, the default language, and keyword recognition.
• Call behavior: It is possible to set options such as silent reasking, timeout for hanging up, background sound, and whether recordings should be saved; these settings allow the experience to be adjusted according to the business risks involved.

Knowledge base and retrieval enhancement

• Data import: Supports PDF, JSON, TXT, CSV, and web content; the maximum size for a single file uploaded is 50MB.
• Category-based management: Separate FAQ, prices, vehicles, products, and policies into different knowledge categories to prevent irrelevant content from interfering with each other.
• Structured queries: CSV allows searching based on specified columns and controlling the fields that are returned, making it suitable for precise inquiries such as vehicle capacity, service prices, and product lists.
• Search control: Chapter Count determines the number of segments returned, while Threshold adjusts the range of relevance; too broad parameters can introduce noise, whereas too strict parameters may result in missing relevant answers.
• Category instructions: Administrators can require that only exact matches be used, a fixed output format be adhered to, or no response be given in the absence of evidence, thereby reducing the likelihood that the generated content deviates from the business data.

Flows, Actions, and business execution

• Visual workflow: Use conditions, variables, and actions to design processes for scheduling appointments, identity verification, order lookup, assessment of lead eligibility, as well as notification and transfer to human agents.
• External APIs and Webhooks: Used to call business interfaces from conversation parameters, receive events, or write results back to the system; in a production environment, it is necessary to handle authentication, timeouts, retries, and idempotency.
• Code actions: Workflows can execute controlled Python functions to carry out calculations, formatting, and branching logic; keys should not be exposed within them, nor should unverified external input be processed.
• Communication functions: It allows sending emails, text messages, and team notifications, as well as handling confirmations, reminders, post-meeting summaries, and escalation requests.
• Knowledge base writing: During a session, tickets, feedback, or leads can be added to a structured CSV file; however, it is necessary to verify the fields, permissions, and for any duplicate records before performing the write operation.

Human agents and operational analysis

• AI Workspace: Provides live suggestions, recommended answers, knowledge search capabilities, as well as information on the customer’s emotions and context, to assist agents in handling complex conversations.
• Call handling: Supports inbound and outbound calls, as well as a hybrid approach combining AI and human assistance; it also provides session summaries and data to the agent who takes over the call.
• Dashboard: Aggregates data on voice and chat interactions, agent activities, sentiment, usage, overages, and AI costs to facilitate process optimization.
• Audio import: Convert existing call audio into text and generate a summary; the balance of the plan is deducted based on the number of minutes in the recording.
• Real-time control: Flow 2.0 includes session monitoring, debugging capabilities, model and action libraries, as well as the option for brand customization; however, the actual scope of these functions may be limited by the specific setup.

Input, output, and channels

Creation and deployment process

1. Register and verify your account using an email address, create an organization, and then establish the first AI agent from a template or a blank project.
2. Set the name, opening statement, system instructions, language, and voice, then check the basic conversations at the Chat and Web Call testing interfaces.
3. Create a classification structure for the knowledge base based on topics, upload documents or import web pages, preview the content, and set parameters such as relevance thresholds, the number of segments to return, and the output fields.
4. Design the processes and actions, integrate with calendars, CRM systems, order management systems, email, SMS, or custom APIs, and assign the minimum required permissions to all keys.
5. Configure manual transfer, prohibited actions, and session auditing for identity verification, payments, medical services, complaints, and incorrect answers.
6. Use the free chat replies and voice minutes provided in the backend for testing; the free credits can only be used within the administration panel, and they cannot be used to publish public components or connect to phone numbers.
7. After selecting a paid package, you can set up web chatting, web calling, dedicated numbers, or SIP lines, and conduct end-to-end testing of the actual communication channels.
8. Continuously monitor the hit rate, transfers, recordings, cost per transaction, and any overcharges, and update the relevant knowledge, prompts, and failure recovery procedures.

Price packages

Pay-per-use rate and package balance

The monthly fee creates an account balance that can be used for various services, and it does not include the amount available for each service listed in the table at the same time. For example, if the $200 allocated to the Starter plan is used entirely for messaging, around 1600 replies can be sent; when using the funds for multiple services, the cost is deducted based on the actual price of each service.

Any usage that exceeds the amount included in the package will be charged according to the excess pricing rules. The Workspace minutes cover only manual call handling; AI suggestions, summaries, or other features used during calls incur additional AI fees.

Integration, API, and open-source status

Suitable for users and scenarios

• Customer service center and BPO: They handle identity verification, order status, common inquiries, and appointments automatically, transferring complex requests to human agents.
• Sales team: Provides 24/7 support, assesses candidate qualifications, gathers leads, schedules demonstrations, and updates external systems.
• Transportation, hotels, and dining: Search for vehicle types, itineraries, available seats, and menus; complete reservations, confirmations, and set reminders.
• E-commerce and retail: answering questions about products, checking orders, handling returns and exchanges, and sending notifications.
• Medical and financial teams: It is possible to create low-risk service processes, but it is necessary first to address issues related to regulation, data retention, consent for recording, and manual review.
• Development and integration team: Uses APIs, Webhooks, Python scripts, and industry-specific connectors to integrate conversations with existing business systems.

Privacy, security, and usage restrictions

• The privacy policy states that the platform processes account, contact, purchase, device, usage data, as well as voice and chat data; calls and conversations may be monitored or recorded for the purposes of fulfilling obligations, ensuring quality, conducting analysis, and resolving disputes.
• Payments are processed through a third party, and the terms state that the website does not store credit card information or other billing data; nevertheless, the company still needs to determine the payment service provider and the location where the data is stored.
• The website footer displays the ISO 27001:2022, PCI DSS 4.0.1, and SOC 2 certifications, but the public pages do not provide a complete overview of the scope of the reports. Purchasers should request valid certificates, audit reports, and a list of the systems in question.
• The option to save voice recordings can be enabled or disabled in the proxy settings; however, the specific retention periods for transcriptions, logs, backups, and analysis data have not yet been made public – these details must be specified in the contract and data processing agreement.
• In medical environments, it is not sufficient to purchase add-ons with no retention options; rather, a signed BAA must be obtained before handling protected health information, and the encryption, access control, and deletion procedures must be verified.
• The terms explicitly prohibit the use of the service for high-risk applications such as public safety, emergency dispatch, life safety, and related communication relaying.
• AI responses may be inaccurate, outdated, or inconsistent; decisions related to accounts, healthcare, finance, qualifications, and payments require traceable human review.
• No refunds or partial refunds are provided for processed payments; closing the account in advance will result in an immediate suspension of services, with no refund for any unused time.

Advantages and limitations

• Agent building, knowledge bases, voice functionality, processes, agent assistance tools, and operational analytics are all integrated in one platform, making it suitable for scaling from pilot projects to hybrid customer service models.
• PDF, CSV, JSON, web pages, and industry connectors cover a variety of data types, but outdated knowledge and improper retrieval settings can still lead to incorrect answers.
• The monthly balance can be used across different projects, which facilitates the combination of voice calls and chat; however, when the costs related to Workspace and AI are added together, it is necessary to monitor them carefully.
• The lowest priced plan costs $200 per month; the free version cannot be used on official websites or for phone calls, and it is not suitable for users who only need a personal chatbot.
• The public privacy policy primarily describes data processing at the website level; customers should consult separately the policies regarding product data, model providers, as well as those related to data storage and training.
• The compliance badge, the level of data retention at zero, and the actual terms of the enterprise SLA must be verified through reports and contracts; they cannot be determined solely based on marketing pages.

Summary

Monobot is suitable for companies that need to develop AI-based voice services, chat functions, and manual support workflows simultaneously; it enables the execution of customer service and sales tasks through knowledge bases, process actions, and various integrations.

During the evaluation, the monthly balance, mixed unit cost, excess amounts, audio and transcription records, manual transfers, compliance evidence, and refund rules should all be taken into account as part of the pilot acceptance process.