Converge Features

It puts every conversation from every connected channel into one list, with full-text search across contacts and messages, so an agent never switches apps to answer a customer.

Each conversation carries the context an agent needs to answer without asking twice: the customer's profile (country, language, email, phone, device, IP), their full history across channels, internal notes from teammates, colour-coded tags, and a page-by-page journey timeline for widget visitors.

• Real-time updates over WebSocket — new messages, typing indicators, and a live indicator showing which teammate is already viewing a conversation.
• Assignment and priority — assign a conversation to an agent and set priority (urgent, high, normal, low), which drives both SLA targets and inbox sort order.
• Status and bulk actions — open or resolved, with bulk status change, bulk tagging and bulk delete.
• Attachments — send and receive images, documents and media, encrypted at rest and served through expiring signed URLs.

The deeper walkthrough lives in the inbox guide, and the category-level background in our unified inbox guide.

Nine channels connect self-serve: the website widget, WhatsApp, Messenger, Instagram, Telegram, Zalo Official Account, Discord, Email over IMAP/SMTP, and X (Twitter) DMs.

LINE, WeChat, KakaoTalk, Viber, LinkedIn, Slack and TikTok are not self-serve — contact us and we will discuss your use case rather than pretend the toggle exists.

Alongside channels, the Shopify connector pushes context rather than messages: an agent sees the shopper's current product view and live cart in the customer sidebar. It runs entirely in Shopify's client-side sandbox and requests no Admin API access, so it cannot read order history.

A brandable chat surface that also works as a lead-capture and self-service layer: visitor beacon tracking, FAQ answers, suggested conversation starters, optional email or phone collection, and page-view analytics.

Installation is a script tag, a Google Tag Manager community template, the WordPress plugin, or the Shopify app — plus written guides for Wix, Squarespace, Webflow, WooCommerce and BigCommerce.

• Before a chat starts: the beacon records page views, UTM parameters and device info, so an inbound message arrives with the browsing path attached.
• Self-service: a Q&A knowledge base answers common questions without an agent, and suggested messages give hesitant visitors a way in.
• Customisation: brand colour, welcome message and popup message; profile collection can be required or optional.
• Abuse control: abusive visitors can be banned from sending messages, which also cleans up their analytics data.

Setup detail is in the widget guide, and the wider category comparison in our chat widget guide.

The AI drafts replies for agents and can answer new conversations directly, but only within limits you set — it answers from your approved knowledge base, marks its own autonomous messages as AI-generated, and hands off to a human when unsure.

Autonomy has three settings. Off gives suggestions only. Conservative replies solely from knowledge-base facts. Balanced, the default for new accounts, also answers confident general questions. Accounts that existed before the feature shipped stay off until they opt in.

• Knowledge base grounding — facts you add manually, that the AI proposes when a conversation closes, or that it distils from an uploaded PDF, Word or text file. Every proposal goes through a human review queue, and the uploaded file itself is never stored.
• Contradiction detection — a new fact that disagrees with a stored one is flagged as a replacement instead of being silently merged; approving the correction retires the outdated fact.
• Guardrails — the AI hands off after 10 consecutive AI replies with no agent involvement, declines off-topic requests, never answers the same customer twice in a burst, and never auto-replies to out-of-office messages or bounces.
• Also included — message translation with company context, quick-reply matching, tone settings (professional, friendly, casual), and BYOK if you prefer your own provider key.

Full behaviour and review flow: the AI guide.

Routing, acknowledgements and reusable replies are all rule-based, so they behave identically every time regardless of what the model does.

• Auto-routing — round robin or load balanced, optionally to online agents only, preferring agents currently on shift, with a log of every assignment decision.
• Auto-replies — handoff acknowledgements and away messages rather than generic greeters, scheduled against working hours, controllable per channel, with A/B variants and thumbs up/down feedback.
• Quick replies — reusable templates with optional attachments, organised into collapsible groups, with dynamic variables (customer name, company, and your own custom key-value pairs) previewed before sending.
• Tags and notes — colour-coded tags in groups for categorisation, and internal per-customer notes with an edit history.

See the replies guide and tags guide for the day-to-day mechanics.

Roles, shift schedules, internal chat, SLA policies, CSAT surveys and per-agent analytics — the pieces you need once more than two people share an inbox.

• Roles — Owner (everything including billing), Admin (employees and company data), Agent (chat only). Login is passwordless via emailed codes, and active sessions can be revoked per company.
• Working hours — company-wide hours and timezone drive auto-replies and SLA maths; per-agent shifts influence routing and let agents silence push notifications off shift.
• SLA policies — first-response and resolution targets per priority, calculated in business hours, with due and overdue badges and breach notifications.
• CSAT — 5-star surveys sent on resolve at the agent's discretion, with company and per-agent scores and a 48-hour privacy embargo on comment text.
• Analytics — messages sent and response times per agent, conversation and lifecycle metrics, weekly response-time tracking.
• Team chat and presence — internal messaging between employees, online presence, and web push notifications on browser and installed PWA.

On the compliance side: per-company envelope encryption (AES-256-GCM) for customer PII, integration credentials, uploads and BYOK keys; a dot-namespaced audit log of sensitive actions; full company data export as a ZIP; and configurable retention for anonymous visitors and their analytics events.

$49/month flat for up to 15 team members, with a 7-day free trial that does not ask for a card. Every feature on this page is in that one plan — there is no higher tier to unlock AI, SLA or channels.

Per-seat competitors bill each agent separately, so the gap widens as you hire: a 5-agent team on a $55/seat plan pays $275/month for the same headcount. See pricing for the full breakdown, or the comparisons for vendor-by-vendor maths.