Four channels, one agent
Your customer writes on the website, in RCS, over SMS or on WhatsApp. The same agent answers, from the same knowledge base and at the same cost. Voice is on the way.
The same agent answers on your website, in RCS, over SMS and on WhatsApp. You set it up once, in one dashboard, on one knowledge base. First conversation in 15 minutes.
You switch a new channel on with a toggle, with no second bot to maintain.
Six things that usually take three projects with three vendors. Here, one panel, one price.
Your customer writes on the website, in RCS, over SMS or on WhatsApp. The same agent answers, from the same knowledge base and at the same cost. Voice is on the way.
Sales, support, bookings — each with its own personality, tone, and tools. All managed from one panel.
Upload a price list PDF, link to your terms, FAQ from help centre. The agent answers with facts from your own materials — never makes things up.
Define typical conversations (“what about returns?”, “how do I file a claim?”). The agent passes them before every deployment.
It scans conversations where it stalled or answered poorly, pinpoints the exact knowledge gap or prompt bug, and rewrites itself. Better every week — with zero effort from you.
Every conversation auto-rated (happy / neutral / frustrated). A chart shows the weekly trend so you know when something needs attention.
The agent recommends real products straight from your Google Merchant feed — live product search, kept in sync on a schedule.
Conversation counts, average length, busiest hours, and volume per channel — one dashboard, no spreadsheets.
No code, no integrator, no weeks of waiting. Open an account in the morning, customers are talking by the evening.
Log in with Google or Microsoft. 30 seconds and you’re in.
↳ No credit card. Cancel anytime.
The website snippet is one line of code. RCS, SMS and WhatsApp switch on in the dashboard, each with its own toggle. The next channel is not a second rollout.
↳ Your IT person will handle this in 5 minutes.
PDFs, a link to your website, FAQ from your help centre. The agent eats it and knows what to answer.
↳ You can also start empty — the agent will manage.
Agent passes your tests, you click Deploy. First real conversation in moments.
↳ Versioning and rollback in one click.
Averages from SMB deployments in 2025. Your results will depend on your industry — every deployment starts with an audit.
Agent answers the usual order-status, refund, and shipping questions. Your team only handles the hard cases.
Customer asks about a product mid-checkout — gets an answer instantly. Fewer abandoned carts, more completed orders.
Not “we’ll get back to you in 24h.” Not “let me check.” Right here, right now — regardless of hour or weekday.
One agent does the work of five support people. No breaks, no holidays, no weekends. Scale without hiring.
* Averages from 47 SMB deployments in Poland, Jan 2025 – Mar 2026. Methodology at docs.chatmerce.eu/roi.
Set up in 12 minutes, connected to a product catalogue and FAQ. Recommendations, stock checks, product cards — all without leaving the conversation.
Same components, different configuration. Each of these runs on the website, in RCS, over SMS and on WhatsApp. Start with whichever fits your business best.
Agent knows your catalogue, stock levels, and order status. Customer asks — agent answers with facts. Fewer abandoned carts, more returning customers.
Transparent, controllable AI for European e-commerce — disclosed by default, human-approved, and processed in the EU. The EU AI Act's transparency rules for AI chat apply from 2 August 2026 — Chatmerce meets them by default, today.
Every chat shows an AI notice up front — and it can't be removed.
AI-proposed configuration changes apply only after a human approves them.
Conversations, knowledge bases, and secrets are processed in the EU — Google Cloud europe-west4 (Netherlands).
API keys stay in Google Secret Manager, attached server-side, never passed through the model or logs.
Admin and agent-configuration actions are logged — who, what, and when.
Inbound visitor messages are scanned with Google Cloud Model Armor.
Always a free plan. Plans tailored to your organisation. LLM (AI model) tokens included in the price.
Web, RCS, SMS and WhatsApp are included in every paid plan and draw from one shared message pool. Carrier delivery of a message is billed separately. All prices net.
Add only what you need — independent of your plan.
Non-expiring. Auto-purchased once the monthly quota is exhausted.
Add another agent without upgrading the whole plan.
Removes the “Powered by Chatmerce” footer from the widget.
Couldn’t find your answer? Write to hello@chatmerce.eu — we reply within one business day.
All conversations, knowledge bases, and secrets are in the European Union — region europe-west4 (Netherlands). Passwords and keys live in Google Secret Manager, never in the database or logs. Full GDPR compliance; DPA available on request. Enterprise plan: detailed sub-processor list.
Web agent only: 5 minutes, drop the snippet and it runs. RCS or SMS in developer mode: 15 minutes. WhatsApp: 15 minutes on our side once you have your own WhatsApp Business sender. Full RCS activation with mobile operators, on real numbers, takes 1 to 2 weeks on SerwerSMS’s side, independent of us.
Yes, anytime from the panel. Upgrade — applies immediately, pro-rated. Downgrade — applies from the next billing period.
All European languages, plus English, Spanish, Italian, and more. Around 100 total. Polish, English, and Norwegian are tested natively. For niche languages we recommend prompt tuning on the Growth or Enterprise plan.
Your call. The agent either asks a clarifying question, or admits it doesn’t know and escalates to a human (with full conversation context). Strategy is set in the agent’s prompt.
Only if you want the RCS or SMS channel. SerwerSMS is our partner and you’ll have a separate contract with them, cheaper than going direct to mobile operators. WhatsApp needs your own WhatsApp Business sender, which you connect in the Chatmerce dashboard. The web agent works fully without either.
Bundle is 50 KB, loaded asynchronously after the page renders. Lighthouse Performance stays above 95 — we tested on 12 Shoper stores and 8 WordPress sites.
We have an API endpoint and a panel button that removes all of a user’s data (conversations, embeddings, identifiers) from the database, BigQuery, and vector index — in one operation. You can expose this to your own customers as self-service.
Sign up, plug in your sources, deploy. No card, no integrator, no annual contract.