InsertChat

What is InsertChat?

InsertChat is an AI assistant for websites and an AI-based phone reception platform designed for businesses. It reads web pages, documents, videos, FAQs, product information, and business rules approved by the company, and uses a unified knowledge base to answer questions from visitors and callers.

The platform can also collect leads, handle reservations, process payments, check orders, create tickets, and transfer complex conversations to human agents. It is suitable for creating an AI front desk for businesses, rather than serving as a general chat model for individuals.

A one-sentence summary

InsertChat integrates website chat, phone support, knowledge base queries, business tools, and human intervention to create a 24/7 AI front desk that can be deployed with a custom brand.

One knowledge base serves two channels.

Companies can opt to use only the website assistant, only the AI receptionist, or both of these tools simultaneously. Business information, services and FAQs, booking rules, calendars, CRM systems, upgrade policies, shared inboxes, and result analysis can all be shared among them.

Main functions

Corporate knowledge base Q&A

InsertChat retrieves answers from approved business materials and can display tables, images, forms, and buttons within conversations. Teams should only connect to content that they have permission to use and that is suitable for presentation to customers.

Website AI assistant

The website assistant can answer questions regarding products, delivery, returns and exchanges, opening hours, and service procedures 24/7, as well as guide visitors through the next steps. Its appearance, avatar, color, and tone can be customized according to the brand’s specifications.

AI phone reception

The paid plan includes AI-powered phone reception along with a phone number that can automatically answer calls, speak, record requests, and arrange appointments. Teams can join real-time conversations, and call forwarding as well as fallback options can be configured.

Clues and CRM

The platform can collect names, contact information, requirements, and scheduling intentions during chats or calls, and save this data in a CRM system or an external system. The fields used for collection should be limited to those that are truly necessary, and visitors should be informed about the purpose of such collection.

Reservation and payment

Assistants can connect to calendars to check times, collect reservation details and determine the next steps; they can also use Stripe to process payments within chats. For high-value transactions, refunds, and dispute resolution, human verification is still required.

Orders, forms, and work orders

The built-in tools allow for querying orders from Shopify and WooCommerce, filling out forms, creating tickets, or initiating external processes. For each of these actions, permissions, confirmation messages, and error handling mechanisms should be in place.

Live Inbox and manual takeover

AI can draft responses, which can then be sent directly by the team, edited, or used to take over the conversation. The shared inbox allows for task assignment, setting priorities, adding team notes, storing previously drafted responses, and enabling automatic escalation.

Dialogue analysis

The platform organizes conversations by theme and sentiment, and provides analyses on satisfaction levels, leads, purchases, team performance, and content gaps. Monthly reports can be sent by email, but companies still need to verify them using actual sales and customer service data.

What types of content is supported by the knowledge base?

The system can read web pages again as scheduled, and some packages allow for automatic updates in case of changes to the content. Before going live, it is necessary to designate a person responsible for managing the data, to establish rules regarding update times, and to put in place mechanisms for withdrawing updates, so as to prevent the assistant from using outdated policies.

Supported operation tools

• Collect and store information on potential customers;
• Check the calendar and schedule meetings or service appointments;
• Collect payments through Stripe during chats;
• Query Shopify or WooCommerce orders;
• Fill out forms, create customer service tickets, and initiate external processes;
• Search for a webpage and load the specified page;
• Send conversation records and trigger automations;
• Upgrade to live chat or phone transfer when needed.

Process for launching a website assistant

1. Create an account and choose Website Assistant as the first channel to use.
2. Paste the company’s website and wait for the system to retrieve the services, products, policies, and FAQs.
3. Review each item of the imported content and remove any materials that are not accessible, have expired, or should not be made public.
4. Set the assistant’s name, avatar, color, tone, welcome message, and response boundaries.
5. Connect to forms, calendars, CRM, tickets, or e-commerce tools.
6. Testing is carried out using common questions, ambiguous questions, sensitive questions, and malicious prompts.
7. Configure manual escalation for responses that cannot be provided, low confidence levels, and high-risk operations.
8. Copy component code, use plugins, or publish separate pages, then monitor actual conversations.

Process for launching telephone reception services

1. Enable the AI receptionist in the paid account and assign a phone number.
2. Fill in the business hours, service area, booking rules, and transfer conditions.
3. Select the voice, speaking style, opening phrase, and the information that needs to be collected.
4. Connects calendars, CRM, and phone business processes that need to be carried out.
5. Test accent, noise, interruptions, silence, repeated questions, and number entry.
6. Configure fallback paths for emergencies, complaints, sensitive operations, and identification failures.
7. Ensure that call recordings, notifications, consent, and data retention comply with local regulations.
8. First, conduct a trial operation over a limited period of time or with a small number of calls, and then gradually expand it.

Manual takeover and upgrade design

A good AI customer service system should not leave all conversations to the algorithm. Companies need to determine which issues can be resolved automatically, which require human intervention, and how to inform customers when no one is available to assist them.

Deployment method

Integrated ecosystem

InsertChat offers built-in connections, an automation platform, Webhooks, REST APIs, MCP, and other methods for integration. The specific features, package requirements, and third-party fees should be checked before activation.

Supported languages and AI models

The platform supports 58 languages, enabling text or voice services for visitors from different regions. The availability of a language does not mean that it offers the same quality in terms of accent recognition, knowledge retrieval, and tool invocation; each language must be tested separately.

InsertChat supports multiple model providers; users can choose a model for their assistant based on quality, speed, and cost, or they can use their own keys. The available models are constantly changing, so it is necessary to keep track of the model version, cost, and rollback settings in a production environment.

Which users is it suitable for?

• E-commerce businesses that need to answer questions regarding delivery, prices, returns and exchanges, as well as product details, on a 24/7 basis;
• Local service businesses that require reservations, telephone reception, and call forwarding;
• Small and medium-sized teams hope to reduce redundant customer service staff and maintain human intervention.
• Customer operations departments that need to manage websites, emails, messages, and phone calls in a unified manner;
• Agents and service providers that deliver white-label AI assistants to multiple clients;
• Development teams that have their own proprietary products and need API, Webhook, or MCP integration;
• Companies that need multilingual customer service and information on cross-regional operations.

In what situations is it not very suitable?

• Only users who need a permanent, free personal chat robot;
• Teams that are unable to organize accurate business information and policies;
• Organizations that require a completely offline setup, local models, and an isolated network for deployment;
• It is hoped that AI can be used to handle high-risk refund, medical, or financial decisions without supervision;
• Projects that aim to collect sensitive customer data without any legal basis;
• Purchasers who cannot accept changes in points, phone minutes, and costs associated with third-party models.

Product advantages

• The website and phone share the same knowledge base and business configurations;
• Start from a website to quickly create testable assistants;
• Supports web pages, documents, videos, FAQs, catalogs, and cloud storage files;
• It includes built-in tools for tasks such as leads, appointments, payments, orders, and tickets;
• A shared inbox enables a seamless transition from AI drafting to human intervention;
• 58 languages are available for cross-regional customer service;
• White-labeling, customer accounts, and multi-assistant capabilities are suitable for agents;
• REST API, Webhook, MCP, and various integration options support expansion.

Usage restrictions

• The quality of the responses depends on the accuracy, completeness, and frequency of updates of the imported data.
• The model may still misinterpret the question, overlook constraints, or produce incorrect answers;
• Complex complaints, exceptional policies, and high-risk transactions require manual handling.
• Third-party models, telephone services, payment options, and integration services have their own policies and fees;
• Monthly points are consumed based on the length of the conversation, as well as by the use of models, data processing, and various tools.
• Speech recognition is affected by accent, noise, the quality of the connection, and proper nouns.
• The package imposes restrictions on assistants, data, seats, file size, and advanced features;
• The service is provided as it is, without any guarantee of uninterrupted operation or absolute security.

Prices and packages

InsertChat offers a 7-day free trial, with monthly plans starting at $98; opting for an annual payment gives a discount equivalent to two months’ worth of service. All paid plans include a website assistant, an AI receptionist, and a phone number, and the main differences lie in capacity, white-labeling options, integration capabilities, and operational controls.

Phone numbers and additional numbers

The additional phone numbers listed by Growth and Scale cost $10 per month each; the pricing for enterprise deployments is determined based on the specific plan. In addition to the cost of these numbers, calls also consume voice credits as well as possible fees imposed by the service provider.

How is the score calculated?

The cost of website conversations depends on their length, the model chosen, the data imported, and the tools used; as a rough estimate, 10 points are required for each conversation. For AI calls, 12 points are charged per billing minute, which is equivalent to a cost of $0.12 per minute.

Prices, points, and benefits may change; it is necessary to verify once again the currency used, taxes, the total annual cost, the rules regarding excess usage, and automatic renewal before completing the payment. The current terms consider purchases to be non-refundable, and cancellations generally take effect at the end of the paid period.

White-labeling and agent capabilities

The terms of service impose restrictions on commercial resale and customer deployment; agents should choose subscriptions that permit resale and inform end users of the applicable terms. Do not assume that unlimited resale rights are granted just because a white-label feature is available.

Privacy and data processing

The materials uploaded by enterprises remain the property of those enterprises. InsertChat declares that it does not sell personal data, nor does it allow third-party models to use customer content for training purposes. To generate responses, prompts, conversations, uploaded files, and other relevant data fragments may be transmitted via the model gateway to the selected model provider.

Safety and compliance

The platform states that it has undergone a SOC 2 Type II control audit, and its products are described as complying with the GDPR and CCPA requirements as well as being suitable for use in HIPAA environments. Being HIPAA-ready does not mean that all accounts will automatically meet the requirements; medical applications require specific corporate solutions, security audits, and appropriate contracts.

• Data is encrypted both during transmission and when stored statically;
• Role-based permissions determine who can manage materials and workspaces;
• Activities, points, and session logs are used for auditing;
• The data and session data provide paths for deletion;
• Companies can review the zero-model data retention settings;
• Large-scale deployment enables DPA, security, and procurement assessments;
• Any electronic transmission and storage still cannot guarantee absolute security.

API, Webhook, and MCP

InsertChat offers a REST API that enables the creation of assistants, integration with corporate content, management of guest sessions, and invocation of responses based on approved materials. Access to this API is included as part of the Growth package; the Scale and Enterprise packages provide additional features such as Webhooks and higher capacity.

Pre-launch checklist

• Confirm the ownership, validity period, and scope of application for each piece of knowledge material;
• Provide manual escalation for issues related to prices, refunds, complaints, and sensitive matters;
• Testing reveals issues such as injection attacks, unauthorized use of tools, and data leakage.
• Check the website, mobile versions, phone noise, and multilingual experience;
• Configure roles, as well as set policies for deletion, logging, export, and retention;
• Verify model providers, telephone service providers, payment processors, and other sub-processors;
• Set up alerts for monthly points, model billing, and phone minutes;
• Record the rate of manual intervention, the resolution rate, conversion rates, and incorrect answers;
• Re-test high-frequency issues after the policy update.

Basic information

Frequently Asked Questions

Is programming required for InsertChat?

The basic website assistant allows one to start by pasting a website, and it can be published using component code or platform integrations – no programming skills are required. However, for more complex operations, as well as API usage and custom data synchronization, development is still necessary.

Is it possible to have both a website and phone lines at the same time?

Yes. The paid plan includes two channels and one phone number, which allow them to share a knowledge base, calendar, CRM, upgrade rules, and analytics.

How many languages are supported?

The current annotation system supports 58 languages. Companies should use actual business scenarios to test text retrieval, speech recognition, accent handling, and proper nouns.

How long is the trial period?

The self-service trial lasts for 7 days, allowing you to test the AI-based front desk process before making a payment. The specific points, phone number limitations, and restrictions on advanced integrations are specified on the registration page.

How much is the lowest package?

The monthly Pro plan costs 98 dollars and includes 3 assistants, 1000 data entries, 3 seats, as well as 5000 points per month. The annual payment option offers a discount equivalent to 2 months’ worth of service.

Can the InsertChat identifier be removed?

Pro already includes the removal of the platform identifier, while Growth offers more comprehensive white-labeling options, custom domain names, and branded email addresses. Resellers must also have the appropriate subscription license to do so.

How are phone calls billed?

Each paid plan includes one phone number; additional numbers are available at a cost of $10 per month in some packages. AI calls consume 12 points per billing minute.

Are customer data used to train large models?

The platform states that it will not allow third-party foundation models to use customer content for training their models. When generating responses, necessary prompts, conversation details, and text snippets are still sent to the selected model provider.

Is InsertChat open source?

No. It offers development tools such as APIs, Webhooks, and MCP, but its core platform is a commercial hosted service; the complete source code is not made available publicly.

Summary

InsertChat is suitable for businesses that wish to use a knowledge base to serve both website visitors and callers over the phone. With features such as Q&A systems, action tools, a shared inbox, human intervention options, and brand customization capabilities, it represents a more comprehensive AI front-end than a simple chat component.

The quality of the service depends on data management, tool permissions, manual upgrades, and ongoing testing – not just on the choice of model. Before making a purchase, it is necessary to calculate the costs associated with calls based on the actual volume of conversations, and to review the terms related to packages, privacy, resale, and third-party services.