Good questions.
Clear answers.

iRPLY is an AI chatbot SaaS platform that lets you deploy chatbots under your own brand — with full control over look, feel, and customer experience.

iRPLY is designed for online stores, service businesses (clinics, salons, consultancies), digital agencies, and enterprise customers who want to automate customer support and boost sales with AI chatbots.

iRPLY chatbots can automate customer support by answering common questions, recommend products based on customer needs, and provide 24/7 assistance in multiple languages.

iRPLY is designed to handle enquiries outside your team's working hours, including overnight and weekends. Availability depends on the platform, connected services and sufficient intelligence credits; this is not a guarantee of uninterrupted service. Human-agent availability follows your team's hours.

iRPLY's AI automatically detects and responds in the customer's language, including English, Malay, Chinese, and many other languages. No manual configuration is needed.

iRPLY combines business-controlled knowledge, multilingual conversations and configured human handover with unlimited agents and contacts. Usage is governed by intelligence credits—not per-agent or per-contact licences. Customer-facing branding and any additional console branding or BYOM capabilities follow your agreed package.

Contact hello@i-rply.com to arrange a demo and confirm your package, onboarding and account access. The standard initial commitment is 24 months unless your signed agreement states otherwise. Monthly billing is a payment schedule, not monthly cancellation.

In the customer console, open Widgets and choose Create New Widget. Name your chatbot, select or create a skill, and configure its appearance. Add the relevant knowledge, sync it and test the responses before publishing.

Use the installation snippet generated for your widget in the customer console. Add it to your website using the method supported by your website platform. iRPLY can confirm the installation requirements during onboarding.

Yes, you can use the hosted chatbot URL to share your chatbot directly with customers via a link. For website embedding, a simple script tag is all that's needed — no complex coding required.

After creating a widget, you can test it using the hosted chatbot URL before embedding it on your website. This lets you chat with your bot, verify its responses, and fine-tune its skill prompt and knowledge base without affecting your live site.

Go to Widgets > Create New Widget. Fill in the widget name (required) and an optional description. Select at least one skill from the list — you must create a skill first if none exist. Optionally choose an AI personality preset (default is Balanced). Click 'Create Widget' and you'll be redirected to the API Keys tab with your public API key ready for deployment.

Open your widget in the customer console and choose Connect. Copy its generated embed snippet and add it to your website's HTML before the closing </body> tag, or through the equivalent custom-code setting in your website platform. Follow the supplied snippet rather than inventing API method names.

Besides embedding via the SDK script, you can share the hosted Chatbox Link directly (e.g., via email or social media) or download a QR Code (available as PNG or JPG) to print on materials, signage, or packaging.

Create a widget, set up a Product Catalogue knowledge base with your products (via CSV import or manual entry), sync your knowledge base, and attach it to a skill using the Product Advisor preset. The AI will then recommend products when customers ask. Add an FAQ knowledge base for common support questions.

Create a widget, build a Document or FAQ knowledge base with your support information, sync your knowledge base, and attach it to a skill using the Customer Service preset. The AI will answer questions based on your knowledge base content.

A Widget is your configured chatbot instance. Each widget has its own name, appearance settings, skill configuration, AI personality, and deployment settings. You can create multiple widgets for different use cases.

Yes, you can create multiple widgets (chatbots) within your account. Each can have different skills, personalities, knowledge bases, and branding for different purposes or departments.

In the widget settings, you can customize colors (primary, background, text), upload your logo, set a brand name, configure the welcome message, and adjust the overall theme to match your brand identity.

You can apply your logo, colours and brand name to the customer-facing chatbot using the branding controls included in your agreed package. Rebranding the customer console is a separate capability and must be confirmed in your agreement. Branding does not grant reseller rights.

Supported deployment options include an inline chatbot within a page, a floating chat bubble, and a hosted chatbot link. Use the generated installation instructions for the mode you choose.

Yes, you can configure domain binding to restrict your chatbot to specific domains, enforce HTTPS-only access, and set up IP range restrictions for additional security.

Yes, each widget can have a custom welcome message that appears when a visitor first opens the chatbot. You can set this in the widget configuration alongside other appearance settings.

Yes, create separate widgets for each department (e.g., sales, support, billing). Each widget can have its own skill, knowledge bases, AI personality, and branding.

Each widget has a hosted URL where the chatbot can be accessed directly without embedding it on your website. This is useful for sharing via links, QR codes, or in environments where SDK embedding isn't possible.

Yes, you can disable the hosted URL in widget settings so that the chatbot can only be accessed through your website's embedded SDK. This prevents direct access to the chatbot interface.

The Widget SDK is the JavaScript library used to display your chatbot on a website. It supports embedded chat, floating chat bubbles and visitor identification. Use the current snippet and documentation supplied with your widget for the correct setup and method names.

Each widget can generate a QR code that links to its hosted chatbot URL. You can download the QR code as PNG or JPG and use it on printed materials, signage, packaging, or business cards.

The iRPLY chatbot widget works in all modern browsers including Chrome, Firefox, Safari, and Edge. The SDK uses standard web technologies for broad compatibility.

Yes. The chatbot widget supports responsive layouts for mobile screens. Test your chosen deployment on the devices and browsers your customers use, including the surrounding website layout.

A Skill defines your chatbot's behavior and knowledge. It includes a custom system prompt, attached knowledge bases (products, FAQs, documents), and optional plugin integrations that extend the chatbot's capabilities.

Go to the Skills page in the console and click 'Create New Skill'. Write a system prompt that defines your chatbot's personality and instructions, then attach relevant knowledge bases and configure any plugins.

iRPLY provides two built-in skill presets: Customer Service (for support-focused chatbots) and Product Advisor (for product recommendation chatbots). You can use these as starting points, or create a fully custom skill from scratch.

Yes, a single skill can reference multiple knowledge bases of different types — product catalogues, FAQ collections, and document knowledge bases. The AI will search across all attached knowledge bases.

iRPLY offers 9 personality presets: Professional, Concise, Balanced, Friendly, Advisor, Enthusiastic, Creative, Empathetic, and Witty. You can also fine-tune creativity, tone, and response length with sliders.

Write a clear prompt that defines the AI's personality, tone, and behavior rules. Be specific about what the chatbot should and shouldn't do. The prompt is the main instruction the AI follows in every conversation.

The AI interprets the question, selects relevant knowledge or configured tools, retrieves information and generates a conversational response. Processing time varies with the request, model and connected services.

AI can still produce incorrect or unsupported answers. iRPLY uses your knowledge base, configured instructions and tool results to reduce that risk, but it cannot guarantee zero errors. Keep information current, test important scenarios and configure escalation for questions that need human judgement.

Yes. With a Product Catalogue attached to its skill, the chatbot can retrieve products relevant to the customer's needs and present available descriptions, prices and images. Recommendation quality depends on the accuracy and completeness of your catalogue.

Yes, the AI automatically detects the customer's language and responds in the same language. It supports English, Malay, Chinese, and many other languages without any additional configuration.

Configure the chatbot to acknowledge missing information and direct the customer to an appropriate support route or human handover. This helps reduce unsupported answers, but you should still test and monitor how it responds.

Your skill instructions define the topics the chatbot should handle and how it should redirect unrelated questions. Scope controls help keep conversations relevant, but should be tested rather than treated as an absolute guarantee.

Yes, when a compatible helpdesk integration or support route is configured. The chatbot can offer escalation for unresolved or complex enquiries and pass available context to your team. Human responses depend on your team's availability.

Review and update the relevant knowledge base content, then click 'Sync' to refresh the AI's knowledge. You can also refine the skill prompt to give clearer instructions on specific topics.

Response time varies with the AI model, the question, knowledge searches and external tools. Many interactions complete within a few seconds; more involved requests can take longer. A thinking indicator helps visitors understand that processing is underway.

Yes, the AI uses Markdown formatting including bold text, bullet lists, numbered lists, links, and code blocks. Product recommendations are displayed with images, prices, and action buttons.

When the AI is searching your knowledge base or using tools, the chatbot displays a thinking indicator that shows what it's doing (e.g., 'Looking for red shoes...'). This keeps visitors informed.

Bring Your Own Model (BYOM), where included in your agreed package, lets you connect a supported provider account for the user-facing chat model. Auxiliary processing remains on the managed platform infrastructure. You remain responsible for your provider's charges, terms and data-handling arrangements.

Conversational data collection means gathering structured information through a conversation instead of a conventional form. Advanced structured data-collection functionality is not promised as a live feature unless confirmed for your deployment. Discuss your required fields and workflow with iRPLY before relying on it.

Skills define your chatbot's behavior — they contain the system prompt, personality settings, and tool configurations. Knowledge bases contain the information your chatbot can access. A skill references one or more knowledge bases.

A Knowledge Base is a collection of information that your chatbot uses to answer questions. iRPLY supports three types: Product Catalogues (structured product data), FAQ Collections (Q&A pairs), and Document Knowledge Bases (uploaded files).

iRPLY supports three types: (1) Product Catalogues for structured product data with prices and images, (2) FAQ Collections for question-answer pairs, and (3) Document Knowledge Bases for uploaded files like PDFs, Word documents, and spreadsheets.

You can add products individually through the console UI or use CSV bulk import for large catalogues. Each product can include name, description, price, category, image URL, and LLM-optimized fields like highlights and use cases.

You can add FAQ items individually through the console or use CSV bulk import. Each FAQ needs a question and answer, with optional category, tags, and search keywords for better retrieval.

Supported formats include PDF, DOCX, XLSX, HTML, TXT, MD, CSV and JSON. Files are processed into searchable content for retrieval. Confirm the current file-size and spreadsheet-row limits for your account before uploading large datasets.

After adding or updating products, FAQs, or documents in your knowledge base, you need to click 'Sync' to make them searchable by the AI. Only changed items are re-processed, making it efficient.

When a customer asks a question, the AI uses intelligent multi-dimensional search to find the most relevant content across all your attached knowledge bases — even if the exact words don't match.

When you click 'Sync', only items that have been added or modified since the last sync are processed. The system uses intelligent change detection for efficiency even with large knowledge bases.

Yes, a knowledge base can be attached to multiple skills, and those skills can be assigned to different widgets. Updates to the knowledge base (followed by sync) are reflected across all chatbots that use it.

Yes, you can export a CSV template, fill it with your product data, and import it to bulk-add or update products in your catalogue.

Yes, you can download a CSV template, populate it with your FAQ data (question, answer, category, tags, search keywords), and import it to add multiple FAQs at once.

Fill in the additional product fields: highlights (key selling points), use cases (when to use), target audience (who it's for), and search keywords. These fields significantly improve the AI's ability to match products to customer queries.

Products have optional fields designed to improve AI accuracy: highlights, use cases, target audience, search keywords, and LLM context. Filling these in helps the AI make more relevant recommendations.

The AI uses FAQ Q&A pairs as reference knowledge when answering customer questions. It synthesizes the information and responds in its own conversational style rather than parroting answers verbatim.

Documents go through three stages: (1) Secure upload, (2) Automatic processing to extract text and split it into searchable sections, (3) Sync to make the content searchable by the AI.

You can edit knowledge base items while your chatbot is operating, then sync the changes to refresh searchable content. Verify important updates after syncing. Service availability remains subject to your agreement rather than an unconditional zero-downtime promise.

Yes, for Product Catalogue and FAQ knowledge bases, you can export all items as CSV. This is useful for backups, migrating, or editing data externally before re-importing.

The website crawler lets you automatically import content from your website into your knowledge base. It discovers URLs (up to 200), then you select which pages to crawl (up to 100). The crawler extracts products and FAQs using AI.

Crawling consumes intelligence credits for processing the selected pages. The amount can vary with content and processing requirements, so check the current usage estimate or confirm it with iRPLY before a large crawl. Do not assume a fixed per-page rate.

Yes, the crawler uses a full Chromium browser engine to render pages, including JavaScript-heavy single-page applications and dynamically loaded content.