Search our knowledge base, or browse by topic below. Can't find what you need? Our team replies within one business day.
Try a different term, or contact our team directly.
Wrennon is an AI support agent built specifically for ecommerce. Instead of starting from a generic chatbot template, it syncs directly with your product catalog, live order data, and store policies, then handles real customer conversations across chat, email, and WhatsApp.
It can look up an order, explain a return policy, recommend a size, or process a straightforward refund on its own — and it hands off anything sensitive, ambiguous, or outside its configured limits to a human teammate, with the full conversation history attached.
From Settings → Store Connection, choose your platform:
/v1/catalog and /v1/orders REST endpoints, or push data via webhook. See the Integrations section below for API details.Initial catalog sync usually finishes in under 10 minutes for stores with fewer than 50,000 SKUs. You'll see a progress indicator in the dashboard, and the agent won't go live on your storefront until sync completes.
After connecting your store, the setup wizard walks through four steps: (1) confirm your brand name and support hours, (2) pick a tone preset (Friendly, Professional, or Concise — all fully editable later), (3) set your top-level escalation rule (a safe default is provided), and (4) preview the widget on a sample conversation.
You can publish the widget to your storefront immediately after the preview step, or keep it in test mode and share a private link with your team first.
Product data, variants, and inventory sync automatically once your store is connected — no manual entry needed. Policies work differently: add your return window, shipping rules, and any store-specific exceptions under Knowledge Base → Policies, either by pasting existing policy text or answering a short guided questionnaire.
The agent treats anything in your Knowledge Base as ground truth and will not improvise outside it — if a question falls outside what's been documented, it escalates rather than guessing.
Go to Team → Invite Member, enter their email, and choose a role (Agent or Manager). Invited teammates receive a signup link by email and land directly in the agent dashboard once they set a password.
There's no limit on the number of human agents you can add on any plan — billing is based on AI-handled conversation volume, not seat count.
Every plan includes a monthly allotment of AI-resolved conversations — a conversation counts once, regardless of how many messages it takes to resolve. Conversations that are escalated to and finished by a human agent don't count against your AI usage limit.
You can track usage in real time under Billing → Usage, and we'll email you at 80% and 100% of your monthly allotment. Going over doesn't shut off your agent — overage is billed at your plan's per-conversation rate, or you can set a hard cap in Settings.
We bill per resolved conversation rather than per seat or per message, since that's what actually reflects the value delivered. A conversation is counted the moment the AI marks it resolved or a human agent closes it out — reopened conversations within 24 hours don't count as a second conversation.
Plan changes are self-serve from Billing → Plan. Upgrades apply immediately with prorated billing for the current cycle. Downgrades and cancellations take effect at the end of your current billing period, so you keep access to your current plan's features until then. Cancelling doesn't delete your data — catalog sync, conversation history, and settings are retained for 90 days in case you reactivate.
All invoices are available under Billing → Invoices as downloadable PDFs, and are also emailed to your billing contact automatically. We accept all major credit cards and, on annual Enterprise plans, wire transfer or ACH. To update your card or billing email, go to Billing → Payment Method.
Under Agent → Personality, you can adjust tone along a few sliders (formal ↔ casual, concise ↔ detailed, reserved ↔ expressive with emoji), give the agent a name, and write custom instructions in plain English — for example, "never use exclamation points" or "always mention our loyalty program when a customer asks about returns."
Changes apply to new conversations within about a minute; conversations already in progress finish with the tone settings that were active when they started.
Escalation rules live under Agent → Escalation. You can trigger a handoff to a human agent based on any combination of:
When a handoff triggers, the assigned agent receives the full conversation transcript, order context, and a short AI-written summary of what's already been tried — so the customer never has to repeat themselves.
The Knowledge Base (Agent → Knowledge Base) stores everything the agent is allowed to state as fact: policies, FAQs, brand guidelines, and product care instructions. You can add entries manually, import an existing help center via URL, or upload PDFs and the agent will extract structured Q&A pairs automatically for you to review before publishing.
Wrennon automatically detects the customer's language from their first message and replies in kind — no separate setup required for the 30+ supported languages. If you'd rather restrict the agent to specific languages, you can set an allow-list under Agent → Languages, and any conversation outside that list will escalate to a human by default.
Use Agent → Test Mode to chat with your configured agent exactly as a customer would, without it appearing on your live storefront. Test conversations are flagged internally and excluded from your billed usage, so you can iterate freely before publishing.
Shopify and WooCommerce connect in a few minutes through Settings → Store Connection (see "Connecting your store" above for the full walkthrough). Magento connects via the Wrennon Magento 2 extension, available on the Magento Marketplace — install it, then paste the generated store token into your Wrennon dashboard to start the sync.
Go to Channels → WhatsApp and connect your Meta Business account — Wrennon uses the official WhatsApp Business Platform API, so your existing verified number and green checkmark carry over. Once connected, the same agent, knowledge base, and escalation rules you use for web chat apply automatically to WhatsApp conversations.
Note that Meta requires pre-approved message templates for any message sent outside a 24-hour customer-initiated window (e.g. shipping updates); we provide a starter template library you can submit for approval directly from the dashboard.
Connect a support inbox (e.g. support@yourstore.com) under Channels → Email by forwarding or via direct IMAP/OAuth connection for Gmail and Outlook. Instagram DMs connect through the same Meta Business account used for WhatsApp, under Channels → Instagram.
The REST API (api.wrennon.com/v1) exposes catalog, order, and conversation resources for stores on a custom platform or with a bespoke internal tooling setup. Generate a scoped API key under Settings → API Keys. Webhooks can be configured under Settings → Webhooks to notify your systems on events like conversation.resolved, conversation.escalated, or refund.issued.
If your team already works out of Zendesk, connect it under Channels → Zendesk to have escalated conversations create a ticket automatically, with the AI's summary and full transcript included. The Slack integration (Channels → Slack) posts a real-time notification to a channel of your choice whenever a conversation is escalated, so nothing sits unattended.
Once your store is connected, order and shipping data syncs continuously — typically within seconds of an update in your storefront platform. When a customer references an order (by number, or by identifying themselves via email), the agent looks up live status, carrier, and tracking details directly rather than relying on cached or approximate information.
Under Agent → Refund Policy, set a maximum dollar amount the agent can approve autonomously (a common starting point is $50–$100), plus which reasons qualify for an automatic refund versus requiring human sign-off (e.g. "wrong item shipped" auto-approves, "changed my mind" after 30 days does not).
Anything above your threshold is routed to a human with the refund pre-filled and ready to approve in one click, rather than requiring the agent or the customer to start over.
Conversation data is retained for 24 months by default, or per your account's custom retention setting under Settings → Data Retention. Payment details are never stored by Wrennon directly — the agent references order and payment status through your store platform's own APIs rather than holding card data itself. See our Privacy Policy and the Security section below for more detail.
Managers can export conversation history as CSV or JSON from Analytics → Export, filtered by date range, channel, or resolution status. Bulk exports for full account migrations or audits can be requested through support and are typically delivered within one business day.
The console has three parts: a conversation list on the left, the active conversation in the center, and contextual order details that surface automatically above the chat whenever an order is mentioned. Messages from the customer, the AI, and human agents are all visible in one unified thread, so nothing is siloed.
My cases shows conversations assigned specifically to you. Needs attention surfaces anything escalated by the AI and not yet claimed by a human, sorted by wait time. All chats gives visibility across the whole team's queue, and All history includes resolved conversations for reference or QA.
Use the reply-type selector next to the message box to switch between a Reply (visible to the customer) and an Internal Note (visible only to your team, shown with a dashed border in the thread). Internal notes are useful for leaving context for the next agent, or documenting why an exception was made.
Agents can view and reply to conversations, and access order details. Managers additionally get access to the admin dashboard: team management, system-wide analytics, agent performance, and the AI configuration itself. Roles can be changed anytime by an existing manager under Team → Manage.
All data is encrypted in transit (TLS 1.2+) and at rest (AES-256). Access to production systems is role-restricted and logged, and we run regular third-party penetration testing. For a full technical breakdown, see the Security page.
Wrennon acts as a data processor under GDPR for the customer conversation and order data you route through it. A signed Data Processing Agreement (DPA) is available on request for any plan, and is included automatically with Enterprise contracts. EU customer data can optionally be pinned to EU-based infrastructure — contact your account manager to enable this.
Every team member can enable 2FA (authenticator app) individually under Account → Security. Managers on Enterprise plans can require 2FA for the whole team, and can additionally configure SAML-based SSO under Settings → SSO for identity providers like Okta, Azure AD, and Google Workspace.
First, check whether the answer came from your Knowledge Base or was inferred — open the conversation in the dashboard and click View Reasoning on the message in question to see its source. If the Knowledge Base entry itself is outdated or wrong, correct it under Agent → Knowledge Base; the fix applies to all future conversations immediately.
If the entry was correct but the agent misapplied it, use the Flag for Review button on the message — this feeds directly into our model tuning process and typically improves similar cases within days, not weeks.
This is almost always one of three things: (1) the agent is still in Test Mode and hasn't been published — check Agent → Status; (2) a Content Security Policy on your site is blocking the script — add wrennon.com to your allowed script sources; or (3) a theme customization removed the embed snippet from your storefront's theme file. Re-copying the snippet from Settings → Install resolves this in most cases.
Check Settings → Store Connection for a sync status indicator — a red or yellow status usually means an expired API token (common after a platform password reset) or a webhook that was disabled on the store's side. Reconnecting the integration re-establishes both automatically. If sync shows healthy but data still looks stale, try a manual "Force Resync" from the same page.
For anything time-sensitive — the widget is fully down, checkout-blocking bugs, or a suspected security issue — email hello@wrennon.com with "URGENT" in the subject line, or use the priority channel available to Enterprise plans. Non-urgent questions are best sent through the contact form and are typically answered within one business day.
Our team is small on purpose — you'll talk to someone who actually knows the product, not a script.