EVERY ANSWER

Installing VegaMind takes under five minutes and requires no technical knowledge. After signing up, you are given a single line of JavaScript — a <script> tag — that you paste before the closing </body> tag of your website. That is the entire installation.

The script tag works on any HTML-based website regardless of how it was built. You do not need to install a plugin, configure an API, or touch your backend. Once the tag is live, VegaMind immediately begins indexing your website content and the chat widget appears for your visitors.

On platforms like WordPress or Shopify, you can paste the tag via the theme editor in about 60 seconds. On Webflow, Wix, or Squarespace, you add it through the custom code settings. If you are ever stuck, our support team walks you through it.

VegaMind is compatible with any website that allows you to add a custom JavaScript snippet. This covers the vast majority of platforms businesses use today.

Confirmed compatible platforms include: WordPress (all themes), Shopify (all themes), Webflow, Wix, Squarespace, Framer, Ghost, Bubble, custom-built HTML/CSS websites, React and Next.js applications, Vue.js apps, landing page builders like Unbounce and Instapage, and any other platform with a custom code injection field.

The widget uses Shadow DOM for CSS isolation, which means it will not conflict with your website's existing styles, fonts, or layout — regardless of how complex or custom your site design is.

From the moment you add the script tag, VegaMind automatically crawls and indexes your website. It reads your pages, services, pricing, FAQs, about sections, product descriptions, and any other publicly accessible content — then stores that content in a knowledge base that the AI retrieves from whenever a visitor asks a question.

This is called auto-website learning and is available on every plan including the free Starter tier. The AI does not guess or generate answers from general knowledge — it only responds based on content found on your site, which keeps answers accurate and on-brand.

Whenever you update your website, you can trigger a manual re-sync from the dashboard to refresh the knowledge base with your latest content. On the Essential plan and above, you can also manually upload additional documents to supplement what the scraper finds.

A custom knowledge base is a manually curated set of information that your AI bot draws from when answering questions. It goes beyond what your website says — you can upload anything the bot should know: product specification sheets, pricing guides, internal FAQs, policies, onboarding documents, terms and conditions, or any information that matters to your customers but may not live on your public website.

Building one is straightforward: from your VegaMind dashboard, navigate to the Knowledge Base section and upload your files. Supported formats include PDF, DOCX, TXT, CSV, and plain text. You can also paste content directly into the editor. VegaMind processes and indexes the documents automatically — there is no tagging, categorising, or formatting required.

Custom knowledge base access is available from the Essential plan ($15/month) and above.

Your VegaMind dashboard includes a Sync button that triggers an immediate re-crawl of your website. When clicked, VegaMind re-indexes all pages and updates the knowledge base with any new or changed content. This is a one-click process and typically completes within a few minutes depending on the size of your site.

This matters practically: if you update your pricing page, launch a new service, publish a new product, or change your opening hours, you should run a sync afterward so the bot reflects the change immediately. Without a sync, the bot will continue answering based on the last indexed version of your content.

For manually uploaded documents in the knowledge base, you simply re-upload the updated file or edit the content directly in the dashboard — no sync needed for those, as they update immediately on save.

Yes. VegaMind is a bilingual product built from day one for both English and Arabic markets, reflecting its origins as a product of QBS Global, based in Dubai, UAE. The chat widget detects the visitor's browser language preference and can respond naturally in Arabic with proper right-to-left text rendering.

The knowledge base indexer also processes Arabic-language content — so if your website has Arabic pages, or if you upload Arabic documents to the custom knowledge base, the AI can answer Arabic questions using that content accurately. You do not need to build separate bots for English and Arabic; a single VegaMind bot handles both languages within one conversation.

Arabic support is included on all plans, including the free Starter tier.

The Essential plan is currently priced at $15/month as an early adopter rate — a permanent price lock for anyone who subscribes before July 2026. In July 2026, the Essential plan price for new subscribers will increase to $20/month.

Crucially, the price lock is permanent for existing subscribers. If you sign up at $15/month today, that is your price for as long as you remain subscribed — you will never be automatically migrated to the higher rate. This is not a promotional discount with an expiry date; it is a grandfathered rate that stays with your account.

The Professional and Enterprise plan prices are not subject to this increase. All prices are listed in USD and billed monthly. Annual billing options may be introduced in the future — we will notify existing subscribers if and when that becomes available.

When your bot reaches its monthly response cap, it pauses AI responses for the remainder of that billing period. The widget itself remains visible on your site — visitors can still open it and see a message — but the AI will not generate new responses until the limit resets at the start of your next billing cycle.

You will not be charged automatically for overages. You can, however, upgrade your plan at any time mid-cycle to immediately restore full functionality. The monthly limits are: 250 responses on the Starter plan, 1,000 responses on Essential, and 10,000 responses on Professional (Fair Use policy applies). Enterprise plans have custom limits negotiated as part of the contract.

Your dashboard shows real-time usage so you can monitor consumption and upgrade proactively before hitting the cap.

Yes, you can cancel your subscription at any time from your dashboard settings — no phone calls, no cancellation forms, no waiting. When you cancel, you retain access to your paid plan features until the end of the current billing period. After that, your account automatically reverts to the free Starter plan rather than being deleted, so your bot and conversation history remain accessible.

For the paid plan trials (Essential and Professional each have a 14-day free trial with no credit card required), you can cancel before day 14 without being charged anything. If you forget and get charged, contact us within 48 hours and we will issue a full refund — no questions asked.

Beyond the trial window, we do not offer refunds for partial billing periods, but you can cancel at any time to stop future charges.

VegaMind is built with data privacy as a default, not an afterthought. All tenant data is stored in isolated environments with row-level security (RLS) enforced at the database level — no tenant can access another tenant's conversations, leads, or settings. VegaMind does not sell visitor data, does not use conversation data to train shared AI models, and does not share data with third parties beyond what is required to operate the service.

For GDPR purposes: conversation data is stored per-session and accessible only by the business account holder. You can delete individual sessions or purge all data from your dashboard. We do not collect personally identifiable information beyond what the visitor voluntarily provides in the lead capture form.

If you need a Data Processing Agreement (DPA) for enterprise compliance requirements, contact [email protected].

Bot and website limits depend on your plan. The Starter and Essential plans each support 1 bot for 1 website — designed for businesses with a single primary site. The Professional plan supports up to 3 bots across 3 separate websites, making it suitable for agencies managing multiple clients or businesses with distinct sub-brands.

Each bot has its own independent knowledge base, conversation history, lead capture, and analytics — they are fully separated. You can configure each bot differently: different personalities, different widget designs, different notification routing, and different knowledge bases.

For unlimited bots and websites — including use cases like white-labelling bots for your own clients — the Enterprise plan is the appropriate tier. Contact us to discuss volume pricing and multi-site configurations.

VegaMind's Professional plan ($49/month) includes native CRM sync integrations with four major platforms: HubSpot (contact creation and deal pipeline sync), Salesforce (lead and contact sync), Pipedrive (deal and lead sync), and Zoho CRM (contact and lead sync).

When a visitor submits the lead capture form in your chat widget, VegaMind automatically creates or updates the corresponding record in your CRM — no manual export, no CSV import, no copy-pasting. The sync happens in real time at the moment the lead is captured.

If your CRM is not on this list, the Enterprise plan includes full API access and custom integration capabilities. For specific integration requests, contact [email protected] — we actively expand the integration list based on demand.

Yes. The Professional plan includes calendar booking integrations with Google Calendar and Microsoft Outlook / Office 365. When configured, a visitor can go from asking about your service to booking a discovery call — without ever leaving the chat widget or visiting a separate booking page.

The bot qualifies the lead through conversation, collects the necessary details, and then presents available time slots from your connected calendar. The visitor selects a time, and the booking is created directly in your calendar with the visitor's details populated automatically. Both parties receive a confirmation.

This is particularly effective for service businesses — consultancies, real estate agencies, healthcare providers, and any business where the first step in the sales process is a meeting. It closes the gap between visitor interest and calendar commitment without requiring a human to be present.

From the Essential plan onward, VegaMind fires instant lead notifications the moment a visitor submits the lead capture form in your chat widget. You choose which channels to receive these notifications on — you can enable one, several, or all of them simultaneously.

Available notification channels: Email (to any address you configure), Discord (via webhook to any channel in your server), and Telegram (via bot message to any chat or group). The notification includes the visitor's name, email, phone number if provided, and the full conversation transcript up to that point.

On the Professional plan, notifications also extend to Slack (via webhook) and Microsoft Teams. This allows lead routing directly into your team communication tools — the right person gets pinged immediately, regardless of the hour, without anyone having to check a dashboard.

Human takeover — also called live chat handover — allows you or a team member to step in and take over an active bot conversation manually. It activates in two scenarios: either the visitor explicitly requests to speak to a human, or the bot reaches a point in the conversation where it cannot find a satisfactory answer in the knowledge base and flags the conversation for human review.

When a takeover is triggered, you receive an instant notification (via your configured channels on Essential+) with the conversation transcript. You can then open the conversation in your dashboard and respond directly — the visitor sees a seamless transition to a human agent within the same chat window.

This feature is available from the Essential plan. On Starter, the bot remains fully AI-driven without a formal handover mechanism, though the lead capture form still collects contact details so you can follow up manually.

Pre-defined responses are scripted answers that trigger when a visitor uses a specific phrase or asks a specific question. Unlike AI-generated responses — which are dynamically constructed from your knowledge base — pre-defined responses are fixed text that you write and control exactly.

They are useful for situations where precision is non-negotiable: legal disclaimers, exact pricing quotes, specific contact instructions, or any answer where even a small AI paraphrase could be misleading. For example: if someone asks "What is your returns policy?", you might want the bot to reproduce your exact policy text verbatim rather than an AI summary of it.

Pre-defined responses are available from the Essential plan. They work alongside AI responses — you set specific triggers for scripted answers, and the AI handles everything else.

Yes. Full white-labelling is available on the Enterprise plan. Under a white-label arrangement, all VegaMind branding is removed from the platform — including the dashboard, the widget, email notifications, and any other touchpoints — and replaced with your own brand identity.

This is designed for agencies and software resellers who want to offer AI chat as a service under their own product name. You get VegaMind's full infrastructure, multi-tenancy, and AI capabilities while presenting a completely branded experience to your own clients. Your clients never see the VegaMind name.

The "Powered by VegaMind" badge in the chat widget is removed on the Professional plan and above, even without full white-labelling. Enterprise white-label goes further by also removing branding from the admin and client dashboards.

Custom AI agents are purpose-built AI systems trained on your business's internal data — not just your public website. They are available on the Enterprise plan and are designed and deployed by the VegaMind team in collaboration with you.

What makes them different from the standard chat widget: they operate on internal knowledge bases built from your SOPs, CRM records, internal wikis, historical data, and any other proprietary information. They can be customer-facing or employee-facing — answering internal team questions, generating weekly performance reports from your CRM, processing and classifying incoming documents, or routing tasks between departments automatically.

Multiple agents can be orchestrated together (multi-agent workflows) where one agent hands off to another based on the task. For example: an intake agent qualifies a customer inquiry, passes it to a knowledge agent for a detailed answer, and routes the lead to a CRM sync agent — all without human intervention.

An AI chat widget is a conversational interface embedded on a website that uses artificial intelligence to understand and respond to visitor questions in natural language — rather than following a fixed script of pre-programmed menu options.

Unlike traditional chatbots, which require you to map out every possible conversation path in advance, an AI chat widget interprets the intent behind each message and generates a relevant, contextual response on the fly. It can handle phrasing it has never seen before, follow conversational threads, and answer follow-up questions without losing context.

A well-configured AI chat widget functions as a 24/7 digital team member: answering product questions, qualifying leads, capturing contact details, booking meetings, and escalating to a human when appropriate — all without any human involvement for the routine interactions that make up the vast majority of website visitor conversations.

Retrieval-augmented generation (RAG) is the technique that makes AI assistants accurate for specific businesses rather than just generally knowledgeable. Here is how it works in plain terms.

A standard large language model (LLM) like GPT-4 knows a great deal about the world in general but knows nothing specific about your business — your pricing, your team, your products, or your policies. If you ask it about your company, it will either guess or decline to answer. RAG solves this by adding a retrieval step before the AI generates a response.

When a visitor asks a question, the system first searches your knowledge base for the most relevant passages of content. Those passages are then provided to the AI model as context, and the model generates an answer grounded in that specific information — rather than making something up. The result is an AI that knows your business accurately without requiring you to retrain a model from scratch.

A custom knowledge base for AI is a curated collection of documents, text, and structured information that an AI system is explicitly instructed to draw from when generating responses. Think of it as the AI's reference library — everything it knows about a specific business lives here.

Unlike training a model (which requires massive compute resources and months of work), building a knowledge base is fast and accessible. You upload documents — PDFs, Word files, spreadsheets, plain text — and the system indexes them using vector embeddings, a technique that allows the AI to find semantically relevant passages even when the visitor's question does not use the exact same wording as the document.

The practical outcome is an AI that can answer highly specific questions accurately: "What is included in your Professional plan?", "Do you service the Abu Dhabi region?", "What is your lead time on custom orders?" — all answered precisely because the knowledge base contains the source information. This is what separates a genuinely useful business AI from a generic chatbot.

An AI chatbot captures leads by combining conversational engagement with a built-in contact collection mechanism. The process typically unfolds in three stages.

First, the bot engages the visitor by answering their initial questions — helping them understand a product, service, or pricing detail. This builds enough trust and interest that the visitor is willing to share their contact information. Second, at the right moment in the conversation (configured by you), the bot presents a lead capture form asking for name, email, and optionally phone number or other qualifying fields. Third, once submitted, the lead data is stored in your dashboard, the conversation transcript is saved alongside the contact record, and an instant notification fires to your configured channels.

On the Professional plan, the captured lead is also automatically synced to your CRM and can trigger an automated email sequence — meaning the follow-up happens immediately, even if no human is available at the time of capture.

A rule-based chatbot follows a fixed decision tree. You manually map out every possible conversation path in advance — "if user clicks Option A, show Message 1; if they click Option B, show Message 2." It can only handle the exact questions it was programmed for. Anything outside that map gets a confused non-answer or a "I don't understand" response. Building and maintaining these trees is labour-intensive, and they break constantly as products and policies change.

An AI assistant understands the intent behind a message rather than matching keywords. It can handle questions phrased in infinite variations, follow multi-turn conversations without losing context, interpret ambiguous requests, and generate coherent answers it has never specifically been programmed to give — as long as the relevant information exists in its knowledge base.

For most business use cases today, rule-based chatbots create more frustration than they resolve. AI assistants grounded in a custom knowledge base deliver the experience visitors expect: a conversation, not a menu.

Not entirely — and the honest answer is that the goal should be augmentation, not replacement. AI automation is highly effective at the top of the sales funnel: answering initial product questions, qualifying leads, collecting contact information, booking discovery calls, and routing high-intent visitors to the right human team member. These are repetitive, high-volume tasks that currently consume significant sales team time and happen unevenly across the day.

What AI does not do well — and should not be expected to do — is handle complex negotiations, build long-term relationships, navigate politically sensitive enterprise deals, or read the subtle human signals that experienced salespeople recognise instinctively.

The practical outcome for most businesses is that AI automation handles the first touchpoint reliably and at scale, freeing the sales team to focus exclusively on the qualified conversations that actually require a human. Conversion rates typically improve because no leads fall through the cracks while the team is asleep, in meetings, or at capacity.

Any industry where website visitors arrive with questions that need answering before they convert benefits from an AI chat widget. In practice, the highest-impact use cases tend to cluster around industries where: (a) the buying decision is considered and involves multiple questions, (b) leads arrive outside business hours, or (c) the sales team is too small to respond to every inquiry in real time.

Real estate benefits from automated property enquiries, viewing scheduling, and neighbourhood FAQs at any hour. E-commerce benefits from product comparison help, sizing guides, shipping policy answers, and abandoned-cart re-engagement. SaaS and tech benefit from plan comparisons, feature explanations, and trial activation support. Healthcare and clinics benefit from appointment booking, insurance FAQs, and patient intake. Professional services (law, accounting, consultancy) benefit from initial qualification and discovery call booking.

The common thread: a visitor who gets an answer immediately is dramatically more likely to convert than one who submits a contact form and waits 24 hours for a response.

Before AI chat, businesses handled after-hours leads in one of three ways: they did not (the visitor left and contacted a competitor), they used a contact form (with a 24-48 hour response lag that killed conversion rates), or they paid for overnight staffing (expensive, rarely justified for SMEs).

AI chat widgets change the economics entirely. The bot is live 24 hours a day, 365 days a year. A visitor who arrives at 2 AM with purchasing intent gets their questions answered immediately, has their contact details captured, and potentially books a call for the following morning — all before your team starts work. By the time your team logs in, the lead has already been qualified, the conversation transcript is waiting in the dashboard, the CRM record has been created, and an automated welcome email may already have been sent.

The lead that used to disappear because nobody was awake to answer it now becomes a calendar booking. This is the core problem VegaMind was built to solve — summarised simply as: don't miss the 2AM lead.

Multi-agent AI orchestration is an approach to automation where multiple specialised AI agents work together as a coordinated system, each handling a specific part of a larger workflow — rather than one general agent attempting to do everything.

The analogy is a well-run team: instead of one person handling enquiries, research, data entry, scheduling, and reporting, you have specialists. One agent handles initial customer intake and qualification. Another retrieves detailed information from the knowledge base to answer specific questions. A third syncs the outcome to the CRM. A fourth generates a weekly summary report from all interactions.

Each agent is optimised for its specific task, which makes the overall system more accurate, faster, and easier to maintain than a single monolithic AI trying to handle everything. Multi-agent orchestration is available on VegaMind's Enterprise plan, and is particularly suited to organisations with complex, multi-step workflows that currently require significant human coordination.

Safety and privacy in AI chat depends almost entirely on how the underlying system is architected. The risks exist — but they are manageable with the right technical foundations, and a responsible vendor will make those foundations explicit.

In VegaMind's case, the key safety measures are: tenant isolation (each business's data is separated at the database level via row-level security — one tenant cannot access another's conversations), no cross-tenant model training (your conversation data is not used to train shared models or improve responses for other businesses), and no data selling (visitor information is never shared with third parties beyond what is necessary to operate the service, such as sending a notification to your configured email address).

For industries with specific regulatory requirements (healthcare, finance, legal), you should ensure that any AI vendor you work with can provide explicit data processing agreements and clarify where data is stored geographically. Contact [email protected] to discuss compliance requirements before deploying.