What omnichannel support actually requires: one knowledge base, one customer record and handoff rules across website chat, WhatsApp and voice, with a rollout order, a cost example and the metrics that show it is working.
Omnichannel support means one conversation history across web chat, WhatsApp, voice, and email not four siloed inboxes. Unified AI pulls from the same knowledge base and passes context on handoff. Plan channel rollout starting with customer support chatbot plus voice agent.
Omnichannel customer support means a customer can start a conversation on your website, continue it on WhatsApp and finish it on a phone call without repeating anything, because every channel reads from one knowledge base and writes to one customer record. It is for teams that already answer on two or more channels and have noticed the cost of the gap between them: the caller who has to re-explain, the WhatsApp thread nobody on the phone team can see, the two different answers to the same policy question.
The decision you are making is not "which channels should we be on". It is which system holds the record of each conversation, which knowledge base feeds every channel, and in what order to connect them. This guide covers the difference between multichannel and omnichannel, a channel-by-channel rollout order, how context travels between a website chatbot, WhatsApp and a voice agent, what the setup costs on Botcadence, and what to measure. The omnichannel support feature page covers the platform side.
Multichannel means you can be reached on several channels. Omnichannel means those channels share state. The test is simple: if a customer asks about a delayed order on website chat at 9 am and calls at 2 pm, does the person or agent who answers the call already know about the 9 am chat?
| Multichannel | Omnichannel | |
|---|---|---|
| Knowledge base | One per channel or per team | One shared source, edited once |
| Conversation record | Separate inboxes | One customer timeline across channels |
| Handoff to a person | Customer repeats the issue | Transcript and collected fields travel with the handoff |
| Policy answers | Drift between channels | Same answer on web, WhatsApp and phone |
| Reporting | Per-channel deflection | Repeat-contact rate across channels |
Most companies are multichannel by accident: a chat widget from one vendor, a WhatsApp number on a staff phone, a phone system from a third. The move to omnichannel is mostly a data decision, not a channel decision.
Start where customers already contact you, and in the order that gets a shared record fastest. On Botcadence the three core channels have different lead times, which should shape the rollout order.
| Channel | Typical use | Botcadence setup time | Automate first |
|---|---|---|---|
| Website chat | Pre-sale questions, policy lookups, status checks | About 5 minutes after documents are uploaded | FAQ answers, order or ticket status, lead capture |
| Follow-ups, confirmations, customers on mobile | After Meta verification, usually 24 to 48 hours | Booking confirmations, delivery updates, reopened threads | |
| Inbound voice | Urgent issues, older customers, complex cases | About 5 business days for a managed voice agent | Hours, status, booking, routing with transcript |
| Email, SMS, Instagram, Telegram, Slack | Secondary channels per audience | Varies by channel | Same knowledge base, no new content |
A practical sequence is website chat in week one, WhatsApp as soon as verification clears, then voice once the knowledge base has survived two weeks of real questions. Launching voice first means debugging content and phone handling at the same time.
Shared context is a set of fields, not a feeling. Decide what the record holds before you connect anything. The minimum set that makes a handoff useful:
| Field | Why it matters | Where it comes from |
|---|---|---|
| Identity key | Links a phone number, email and chat session to one customer | Phone number on WhatsApp and voice; email or login on web chat |
| Open issue summary | The next channel starts from the issue, not from zero | Written by the agent at the end of each conversation |
| Last transcript | Lets a person verify what was promised | Stored with the conversation |
| Consent flags | Governs whether you may message on WhatsApp or call back | Captured at opt-in, stored on the record |
| Handoff reason | Tells the human why the AI stopped | Set by the escalation rule that fired |
| System references | Order, ticket or booking IDs | Looked up through the CRM or helpdesk connector |
Botcadence agents on every channel read from the same knowledge base and hand off to a person with the full transcript. Records can be written to a CRM or helpdesk through documented connectors such as HubSpot, Salesforce and Zendesk, or through Zapier and webhooks. The CRM or helpdesk, not the chat tool, should be the system of record; the agents write to it and read from it.
A customer asks the website chatbot whether a coffee machine is covered by warranty. The agent answers from the policy document and captures the phone number and order ID. The customer later messages on WhatsApp about a leak. The WhatsApp agent sees the earlier warranty question and the order ID, opens a return case and writes it to the helpdesk. The customer then calls. The voice agent matches the caller ID to the record, confirms the open case, and transfers to a person with the transcript because return approvals are on the human path. Nobody asked for the order number three times.
Botcadence prices voice separately from chat. The Voice AI plan is $299 a month with 1,500 minutes included and $35 per extra 200 minutes. Web chat and WhatsApp start on the Free plan, with Starter at $39, Growth at $99 and Business at $199 above it; Enterprise is custom. Which chat plan you need depends on reply volume, so check the pricing page for the current allowances.
A worked example, with every input labelled as an assumption. Say a retailer takes 400 support calls a month at an average of 3 minutes: that is 1,200 minutes, inside the 1,500 included on the Voice AI plan, so voice costs $299. Say chat and WhatsApp volume fits the Growth plan at $99. The platform cost is $398 a month before Meta's WhatsApp conversation charges, which vary by country and message category (see the WhatsApp Business API pricing guide). If the same retailer grew to 700 calls at 3 minutes, that is 2,100 minutes, 600 over the allowance, so three $35 packs bring voice to $404 and the total to $503.
Compare that against what the gap between channels costs you now: repeat contacts, the minutes a person spends re-reading history, and the policy answers that had to be corrected.
Choose omnichannel when customers switch channels mid-issue, when policy answers differ by channel today, or when human agents spend time reconstructing history. It is also the right call when you plan to add a voice agent and want it to recognize existing customers.
Stay multichannel, for now, when you have one dominant channel and the others carry little volume, when your system of record has no API or webhook path, or when your knowledge base is not yet consistent. Connecting channels to conflicting content multiplies the inconsistency rather than fixing it.
Botcadence conversation analytics show per-channel volumes and outcomes; the cross-channel repeat-contact rate usually comes from the CRM or helpdesk where the record lives.
Outcomes depend on volume, grounding quality and workflow design, so treat the first month as a measurement period rather than a launch. Botcadence human handoff carries the transcript across every channel, which is the piece most multichannel setups are missing.
If you are starting from website chat, the Botcadence chat agents page covers inbound and proactive chat, the shared knowledge base and the connectors that make it the first step toward one conversation across every channel.
Questions readers ask
No launch website + one messaging app where customers already contact you, then expand.
Multichannel is presence everywhere; omnichannel shares context so customers never repeat their issue.
Usually your CRM or helpdesk ticket bots write transcripts and metadata there, not scattered spreadsheets.
Track repeat-contact rate across channels, not per-channel deflection alone.
Book a free demo and we will build a working AI agent for the channel this article covers, tailored to your business goals. No commitment required.
Ask AI about this page
