chatwize-assist

What is chatwize-assist?

Chatwize-Assist is a pre-release page for an AI customer service solution built using Lovable; the brand name used is ChatWize. It allows companies to upload knowledge bases and deploy website bots, with human customer service taking over when the AI is unable to handle a request.

The current page does not offer official registration, an admin dashboard, or any available customer service chatbots; its main button only opens a form for the waiting list. Therefore, it should be regarded as a page for concept validation and gathering requirements prior to launch, rather than a fully functional SaaS product.

Current operating status

The year at the bottom of the page has been updated to 2026, but the page still asks users to wait for the service to go live. It is not possible to conclude that the service is already in operation based solely on the year indicated, simulated data, or screenshots of its features.

The main features of the plan

Knowledge base training

• The page planning feature supports uploading PDF, DOC, and HTML files, and it also provides a way to access knowledge materials via APIs.
• Companies can assign FAQs, product documentation, and business explanations to AI to answer customers’ repeated questions.
• The page claims to index pages automatically and process them at high speed, but there is currently no operational backend, no limits on file format, and no actual performance tests.
• After going live, it is still necessary to verify that scanning of PDFs, tables, images, links, permission pages, and incremental updates can be handled correctly.

Website chat component

• The demonstration page provides a sample embedding code, showing how the ChatWize component can be integrated into a website using an API key and theme settings.
• Companies can offer round-the-clock Q&A on their websites and customize the chat interfaces to reflect their brand identity.
• The keys, script addresses, and copy buttons displayed are for demonstration purposes only; there are no public documents related to the components, nor any information on versions, permissions, or content security policies.
• For production use, it is necessary to verify cross-domain functionality, script integrity, loading performance, fault tolerance, and key rotation.

WhatsApp and Instagram

• There are plans to integrate WhatsApp and Instagram into a unified inbox, so that AI can handle customer inquiries across multiple channels.
• Multiple channels can help reduce the need for teams to switch between tools, but they still must comply with Meta’s account approval rules, message template requirements, and platform fee regulations.
• At present, there are no actual authorization processes, supported regions, message types, or channel screenshots, so it cannot be determined that the integration is operational.

Clue identification and customer insights

• The page is designed to analyze conversations and identify high-value potential customers, thereby sending alerts to the sales team.
• The insights panel is designed to identify common statistical issues, satisfaction levels, additional sales opportunities, and gaps in the knowledge base.
• The values for session volume, resolution rate, topic proportion, and percentage improvement shown on the page are just examples; they do not indicate information related to actual customers, the time period covered by the sample, or the methods used for calculation.
• Automatic lead scoring can lead to inaccuracies; formal business operations require clear rules, manual review, and mechanisms for handling appeals.

Intelligent manual takeover

• When the AI’s confidence level is below the threshold, it is planned to pass the entire session context to a human customer service agent.
• Such functions prevent the robot from continuing to guess when it is uncertain, and reduce the need for customers to repeat their instructions.
• The page does not disclose the confidence algorithm, queuing, notifications, agent status, service duration, or handling of transfer failures.
• High-risk issues should be escalated compulsorily based on keywords and business rules, rather than relying solely on the model’s self-assessed confidence level.

Planned work process

1. Add yourself to the waiting list, provide your name, email address, and company name, and wait for a notification when the service becomes available.
2. After the product is launched, create an account and set the company brand, the robot’s role, as well as the areas of business for which responses are permitted.
3. Upload FAQs, documents, and product materials, and check that the content is complete, up to date, and that you have the necessary permissions to handle them.
4. Use real customer questions to test responses, declines to answer, and transfers to a human agent; record cases of unanswered questions and incorrect responses.
5. Obtain the official website components and install them in a testing environment to verify performance, security, and mobile display.
6. Use a business channel account when connecting to WhatsApp or Instagram to complete platform authorization and message template approval.
7. Set thresholds for manual intervention, notifications to agents, working hours, and procedures for handling urgent issues.
8. After going live, the conversation and lead data are reviewed, the insights obtained are manually verified, and the knowledge base is continuously updated.

The process outlined above is based on the product design shown on the page; it does not mean that each step can be carried out at this time. Potential users should wait until the official backend system, help documents, and service policies are available before configuring business data.

Input and planned output

Waiting list and pre-sale status

• The candidate registration form collects the name, email address, and company name, so as to notify the applicants once the product is released.
• At the top of the page there is information indicating that a pre-order payment of 20 dollars will grant 6 months of access once the service is launched, but currently there is no payment button or settlement page that can be used to complete the payment.
• There is no public information regarding the delivery date for pre-sales, the features included, the message quota, taxes and fees, refunds, or the procedures for handling project cancellations and delays.
• In the absence of an operating entity and a contract, payments should not be made through informal transfers, private messages, or methods that leave no trace.
• Before submitting candidate information, one should take into account that the page lacks a valid privacy policy and does not indicate how to unsubscribe from marketing emails or have one’s data deleted.

Monthly and annual pricing

The price page shows the planned packages, and all purchase buttons still lead to the waiting list. Starter and Professional offer monthly and annual payment options, while for Enterprise it is only possible to contact sales.

• The pricing page counts every message sent by the customer as well as those sent by the AI within the monthly total; a single round of question and answer typically involves at least two messages.
• Once the limit is exceeded, the page states that a fee is charged per message, but it does not specify the price per message, any maximum amount, or rules regarding suspension.
• On the home page, the simulated dialogues indicate that the Starter plan offers only 500 sessions, while the Growth plan, priced at $79, provides 2500 sessions; this is inconsistent with the information on the pricing page.
• The home page describes Enterprise as offering unlimited support, while the pricing page specifies 10,000 messages per month; it is not yet clear what the final details will be.
• Since the product has not yet been launched, the package names, message definitions, channels, and support responses may change before official settlement.

Free trial and cancellation

Once it is officially launched, it is necessary to confirm the end date of the trial period, whether payment will be activated automatically, the conditions related to message limits, the timing at which cancellation takes effect, and the rules for refunds before submitting a payment request. The options “no credit card required” and “cancellation available at any time” cannot replace a formal contract.

Platform and integration status

Differences from the Chatwize product with the same name

On the Internet, there is also a full-fledged version of Chatwize AI with its own independent domain name; it offers a 30-day trial period, as well as various functions and different policies. The chatwize-assist page does not contain a link to this official site, nor are there any company details indicating that the two services are operated by the same entity.

• The 95 languages, models, customer cases, EU storage options, or 30-day trial period associated with another product cannot be included in this record.
• The address of the mock component on the pre-release page, which includes the same domain name, is not sufficient to prove that the legal entities, accounts, and services have been integrated.
• Users should verify their identity through verifiable company information, contracts, support email addresses, and account access points, in order to avoid confusing two projects with the same name.

API, GitHub, and open-source status

• No official GitHub repository, public source code, or open-source license for chatwize-assist was found.
• The page shows the “Connection API” and the website component keys, but it lacks development documentation, authentication details, endpoints, SDKs, example projects, or rate limits.
• The lovely hosting service and the visible editing logo do not indicate that the project’s source code is publicly available, nor do they grant any rights to copy, deploy, or use it for commercial purposes.
• Before formal integration, it is necessary to obtain the component documentation, API permissions, key rotation procedures, callback verification mechanisms, error handling protocols, and version support policies.

Privacy, security, and commercial risks

The sections on Terms, Privacy, and Cookies on the page contain no actual policy details; information regarding the operating entity, server location, data retention periods, subcontractors, and security measures is not made available.

• The registration form collects the name, email address, and company name, but it does not specify the purpose of processing this data, the retention period, whether marketing emails will be sent, or the methods for sharing and deleting the information.
• Planned customer conversations, documents, leads, and social messages may contain large amounts of personal and business data; real-world operations cannot be undertaken in the absence of appropriate policies.
• The page claims that advanced security, compliance, and SLAs are part of the Enterprise plan, but it does not list the certifications, audit scope, or contractual metrics.
• Automated analysis of conversations and clues requires a legal basis, and it should allow customers to know the identity of the AI, when human intervention is needed, and how their personal information is used.
• The page does not specify the ownership of the knowledge materials, robot responses, and feedback, nor does it address the use in model training, commercial licensing, and boundaries of responsibility.
• There are no guarantee documents for pre-sale deliveries or refunds; companies should not use what is displayed on their websites as a commitment to make a purchase or as proof that the service will be available.

Who should pay attention to it?

• Small businesses that are willing to participate in the preliminary interviews and wait for the service to go live can submit less sensitive candidate information.
• Teams that are comparing AI customer service products can use this page as a list of requirements, rather than considering it as a list of verified suppliers.
• Companies that require immediate deployment, clear SLAs, compliance documents, or verifiable customer cases are not suitable for the current stage.
• Organizations that require APIs, private deployment, data residency, or a formal purchase contract should wait for the technical and legal documents to be made available.
• Any user who plans to submit real customer conversations, internal documents, or payment details for pre-orders should first confirm the operating entity.

Advantages and limitations

Summary

Chatwize-Assist presents a fairly comprehensive vision for AI-based customer service, but as of now it is still in the pilot and proof-of-concept phase. There is no backend interface available for testing purposes, with no support for knowledge bases, website components, multiple communication channels, lead analysis, or manual intervention.

Users who are interested can add it to their watchlist, but the planned price, trial versions, sample metrics, or pre-sale materials should not be considered as features that are already available. It is only after all aspects such as the operating entity, formal contracts, privacy and security measures, refund policies, APIs, and the actual delivery of the product have been confirmed that it is appropriate to proceed with purchasing or integrating the data.