Your Best Support Agent
Never Clocks Out

Aura is Emplorium's AI agent — a configurable, always-on support presence that answers customer questions, handles conversations, and escalates to your human team when it needs to. It's not a decision-tree chatbot. Aura understands natural language, handles follow-up questions in the same thread, and responds based on your actual content — your knowledge base, uploaded documents, and linked URLs. You configure it once, and it works across every channel you're live on.

You feed it your content. Connect your knowledge base articles, upload documents like PDFs and product guides, or add URLs for Aura to crawl. It uses everything you provide as its source of truth — it won't answer from general internet knowledge or make things up outside of what you've given it. The more complete your content, the more accurately Aura answers. You can update its sources at any time and changes take effect immediately.

Yes. Free includes 1 Aura agent, and paid plans include the published number for their tier. You can add more agents as an add-on. Each agent is configured independently — different name, tone, language, escalation rules, and knowledge sources. You might run one agent for general support, another for a specific product line, and a third in Arabic for your Gulf-region customers. They operate in parallel across your channels.

Yes — Arabic is a first-class language, not an afterthought. Aura responds in Arabic natively, with full RTL rendering in the widget. It's available on every plan including Free. You can run agents in Arabic-only, English-only, or configure multilingual agents that detect the customer's language and respond accordingly. We don't charge extra for Arabic or RTL support.

You set the rules. You can configure Aura to escalate based on topic — certain question types it should never answer, like billing disputes or legal queries. You can also set it to escalate when customer sentiment drops, when a conversation exceeds a certain length, or when Aura's confidence in an answer falls below your threshold. Outside of your configured rules, Aura will also escalate on its own if it can't find a reliable answer in your content rather than guessing.

The conversation transfers to your human inbox with full context intact — the customer doesn't have to repeat themselves. Your team picks up exactly where Aura left off. If no agents are available, the customer gets a clear message about expected response time, which you configure. Nothing gets dropped.

Yes. You can define topics Aura should always handle, topics it should always escalate, and topics it should handle only under certain conditions. You can also set explicit do-not-answer rules — for example, never respond to questions about competitor pricing, never discuss refund amounts, always refer legal questions to a specific email address. These rules are applied before Aura generates a response.

Each agent has its own personality settings. You set the communication style — formal, conversational, concise, detailed — and provide a brief about how the agent should represent your brand. You can give it a name, a specific sign-off, and instructions about what to emphasize. You can also provide example responses to calibrate the style. If it drifts, you adjust the instructions and it recalibrates.

Aura routes queries across multiple models depending on response complexity. Simple, factual questions use a lightweight default model — which is why they cost as little as 0.2 credits. More complex, multi-step answers can use a higher-end model for better quality at higher credit consumption. Model selection is automatic today; customer-configurable routing controls are on the roadmap.

Your analytics dashboard shows Aura's resolution rate, escalation rate, average response quality, and credit consumption per conversation. On Pro and above, you get advanced AI performance insights including topic-level breakdown — which question types Aura handles well and which it frequently escalates. You can use this to identify content gaps in your knowledge base and improve Aura's accuracy over time.

Yes. If a customer switches languages mid-conversation, Aura detects the change and responds in the new language. You can configure whether this is allowed or whether Aura should stick to one language per session. Multilingual support is available from the Starter plan.

Aura works across the live chat widget, WhatsApp, Facebook Messenger, and Telegram today. The same agent configuration applies across supported channels. Email, Instagram, and additional channels are on our roadmap.

No. Aura handles concurrent conversations without any queue — there's no "busy" state. Ten customers can message at the same time and each gets an immediate response. The only limit is your monthly credit allowance, which applies to the total volume of responses across all conversations.

Yes. You can preview and test any Aura agent in a sandbox environment before publishing it to your live channels. You can simulate conversations, check how it handles specific questions, and verify escalation behavior. Credits consumed during testing count toward your plan allowance.

Aura can answer order questions from maintained product and policy knowledge today. Live order lookup requires a configured and validated AI data request against an API you control; native store integrations and public webhooks are on our roadmap.

Aura can collect and qualify return requests — including order number, reason, and product condition — and explain your return policy. Actual processing, such as issuing a refund or generating a label, requires a configured and validated Data Update against an API you control; otherwise Aura hands the complete context to a human. Native store connectors and public webhooks are on our roadmap.

Aura can answer product-specific questions based on your catalog content — sizing, compatibility, ingredients, specifications, availability. If you upload product guides or link your product pages as knowledge sources, Aura uses that content to help customers make decisions. It won't hallucinate product details it wasn't given.

Yes. If your current promotions and discount conditions are documented in your knowledge base, Aura can explain them, confirm eligibility, and provide codes. For dynamic discounts that change frequently, the most reliable approach is keeping a short, regularly updated document Aura can reference — it takes effect immediately when you update the source.

Aura absorbs the high-volume, repetitive queries (order status, discount questions, shipping policy) so your agents can focus on the escalations that genuinely need a human. The result is a much smaller support queue during your highest-traffic moments.

Yes. Aura detects the customer's language and responds in kind. For MENA e-commerce businesses, this means Arabic-speaking customers get Arabic responses automatically — including full RTL rendering — without you needing to set up a separate agent or route by language.

Yes, with guardrails you define. You configure how Aura responds to frustrated customers — the tone to use, when to offer an apology, when to fast-track to a human agent. A common setup is to have Aura handle the initial response with empathy and context-gathering, then escalate to a senior agent for anything involving compensation or a formal complaint. The customer never feels dropped.

Direct native integrations are on our roadmap. Today, keep product and policy documentation updated as knowledge sources. If you need live store data, confirm whether a configured AI data request fits your setup before relying on it; public webhooks are not available today.