omnichannel AI customer service

Most "multi-channel" support tools are really N separate bots that share a logo. The Guru answers across five channels from one shared knowledge base, so knowledge and brand voice stay consistent wherever the customer reaches you.

The five channels

The Guru runs on website chat, email, phone, SMS, and WhatsApp. Each channel connects from the admin portal and draws on the same indexed content.

What "shared" actually means

• Knowledge base: one index, queried by all channels. Updating your return policy on the website immediately updates email, phone, and chat answers.
• Brand voice: defined once and applied across channels.
• Escalation rules: a live-agent handoff configured for one line applies the same routing logic when triggered.

Each channel keeps its own conversation history. When a conversation is explicitly moved from one channel to another, that history carries with it: a phone call can offer to continue by text, and a text exchange can invite the customer to continue on WhatsApp, with the Guru picking up where it left off. Your host site can also pass a logged-in customer's context, and a Flow Designer (JSFE) integration can retrieve the customer's records from your CRM to bring relevant context in. Channels are linked when the move is explicit, not guessed at automatically.

A worked cross-channel example

1. Tuesday 2pm: a customer asks via chat: "Can I exchange a hoodie for a different size?"
2. The Guru pulls the return policy and answers, then offers to email exchange instructions.
3. The customer confirms. The Guru sends an email with exchange instructions.
4. Wednesday morning: the customer replies to that email with: "Got the label, can I drop it off at FedEx?"
5. The email channel ingests the reply and the Guru answers based on the same return policy plus the courier details from your knowledge base.

There is no second account, no second bot, no second knowledge base.

Phone is a fresh conversation

Phone calls start fresh every time, by design: each call is its own conversation. When continuity matters on a call, the Guru can offer to continue by text (carrying the context with it), or a Flow Designer integration can look up the customer's records to bring the relevant context in.

Setup model

You enable channels one at a time. Each channel has a connector page in the admin portal (web widget script, email forwarding instructions, Twilio phone number setup, Meta Embedded Login for WhatsApp). The knowledge base is the same regardless of which channels are active. Turn on what you need today; add more later.

The carrier channels follow a Bring Your Own Channels (BYOC) model: for WhatsApp, voice, and SMS, you own the underlying Meta and Twilio accounts directly and pay the carrier directly. instantAIguru never sits between you and the carrier and adds no per-message or per-minute usage markup, only a flat SaaS fee per connected channel.

Trade-offs

• Voice channels add latency (speech recognition + generation + TTS) compared to text, so reply length is tuned shorter to keep calls responsive within sub-300ms response latency in normal conditions.

Why one shared knowledge base matters

Running independent bots per channel produces drift: an updated FAQ that lives only in the chat bot never reaches the email replies. Centralizing the knowledge eliminates that drift and reduces the operational cost of running customer service to one configuration surface instead of six.