WhatsApp Human Handoff Workflow

Outcome Summary

Design an escalation flow where your AI Agent answers what it knows, and creates an Inquiry when it shouldn’t guess.
Keep customers in the same WhatsApp thread while routing “needs a human” chats to the right person with full context.
Improve consistency by turning recurring escalations into better FAQs and instructions over time.

What Clarivo Actually Does (Truth Block)

✅ Clarivo does

Deploys an AI Agent on your WhatsApp Business number to respond to inbound customer messages.
Captures leads automatically when intent is detected (when enabled).
Creates an Inquiry when it can’t answer confidently (when enabled), so it escalates instead of guessing.
Provides a dashboard for conversations, leads, and inquiries—plus an “open in WhatsApp” style human handoff.
Automatically replies in the customer’s language and understands WhatsApp voice notes (with transcription visible in the dashboard).

❌ Clarivo does not

Run outbound broadcast campaigns or mass messaging.
Guarantee appointment booking (it can collect details; humans confirm).
Provide built-in payment processing.
Sync real-time inventory/MLS data.
Act as an omnichannel inbox (it’s WhatsApp-focused).

The Core Problem

Your team needs the speed of an Agent, but not at the cost of the Agent “making up” answers when a customer asks something edge-case.
Even when you do escalate, humans often jump into a chat without context—causing repetitive questions and slower resolution.
Different staff members respond differently, which creates inconsistent policies, pricing explanations, or service qualification.
Voice notes and multilingual chats add friction if your workflow assumes typed, single-language messages.
“Hi” and vague greetings create noise; your workflow needs a clean path from greeting → intent → lead capture or escalation.

Framework

Step 1: Define what must escalate (your “no-guess” policy)

Create a short internal list of situations where the Agent should prefer an Inquiry over improvising, such as:

Policy exceptions (refunds, guarantees, eligibility)
Anything that depends on real-time availability
High-stakes or sensitive questions (complaints, disputes)
Custom quotes or complex requirements

Step 2: Make the Agent strong on the basics (so humans handle only what matters)

Feed Clarivo the inputs it uses to behave predictably:

Business description + operating rules
FAQ (Q/A pairs)
Services/products list (prices optional)
Additional instructions (tone, do/don’t, objection handling, what to ask before quoting)

Step 3: Design your escalation trigger around confidence, not “time”

A clean handoff workflow is usually based on: If the Agent can’t answer confidently, it escalates.

Enable Inquiry creation for “low confidence” situations.
Decide what the Agent should do in the chat when it escalates: acknowledge, set expectation, then gather any missing basics (only if appropriate).

Step 4: Standardize the “context pack” your human needs

Before a human takes over, ensure you can quickly see:

The customer’s last messages and what they’re asking for
Any captured lead fields (service requested, city, urgency, contact info)
Whether the Agent attempted an answer or immediately escalated
The simplest next question that moves the case forward

Step 5: Assign ownership (so escalations don’t stall)

Decide internally:

Who monitors new Inquiries during business hours
Who handles after-hours escalations (or whether you only acknowledge and follow up later)
How you route inquiries by category (sales vs operations vs support)

Step 6: Execute the handoff in the same WhatsApp thread

When an Inquiry is created:

A human opens the exact chat via the handoff flow and replies directly in WhatsApp.
After resolving, the team marks the Inquiry resolved (so the loop is explicit and trackable).

Step 7: Close the loop by turning recurring inquiries into better training inputs

Once a week (or on a regular cadence):

Sort Inquiries by theme.
Promote the best human answers into FAQ entries or operating rules.
Add “what to ask next” instructions for common ambiguous requests.

Copy/paste message templates (edit to your tone)

Escalation acknowledgement (neutral):

“Thanks—this needs a quick check from our team to make sure we answer correctly. I’m escalating it now. A specialist will reply here shortly.”

Clarifying question before handoff (when safe):

“To route this correctly, can you share your city and what service you’re looking for?”

After-hours expectation:

“Thanks for your message. We’ve logged this for the team. We’ll reply here as soon as we’re back online.”

Voice note confirmation:

“Got it—thanks for the voice note. I’m sharing this with our team so they can respond with the right details.”

Use Cases

Use case: Clinic intake (policy + eligibility questions)

Scenario: A patient asks about a service that depends on eligibility and policies.
Recommended approach: Let the Agent handle standard FAQs, then create an Inquiry for edge cases (exceptions, special situations). Human replies in WhatsApp with the final guidance.
Common mistake: Letting the Agent “fill gaps” on policy—this creates risk and inconsistent answers.

Use case: Home services quote requests (lots of missing details)

Scenario: “How much to clean a house?” with no size, location, or timing.
Recommended approach: Agent captures intent + collects a few basics (service type, city, urgency). If the quote depends on conditions, escalate via Inquiry with the captured context.
Common mistake: Quoting too early without required details—leads to backtracking and lower trust.

Use case: Real estate inquiries (availability + specific property questions)

Scenario: A lead asks about a property detail the Agent can’t confirm.
Recommended approach: Agent qualifies the lead (city/area, budget range, timeline), then escalates for anything requiring human verification.
Common mistake: Treating real-time availability as static information.

Decision Checklist

Do you have a written “no-guess” list of topics that must escalate via Inquiry?
Are your FAQs and operating rules complete enough that the Agent can answer the top repetitive questions confidently?
Do you know which lead fields matter for routing (city, service requested, urgency), and are you capturing them consistently?
Do you have a clear owner for monitoring and resolving new Inquiries?
Can your team respond in the same WhatsApp thread quickly using the handoff flow?
Have you defined how to handle voice notes (confirm receipt, summarize, escalate when needed)?
Do you have a feedback loop to convert recurring Inquiries into better FAQ entries and instructions?

Constraints

Clarivo is inbound-focused: it responds when the customer messages first.
It can respond to generic greetings based on configured greeting rules, but it’s not designed for outbound campaigns.
Appointment booking isn’t guaranteed: the Agent can collect preferred times/details; humans confirm.
Payments aren’t handled inside Clarivo.
The Agent can only be as accurate as the information and rules you provide (especially for exceptions).

Common Mistakes

Escalating too late: The Agent tries to “be helpful” on edge cases, which can create incorrect answers and rework.
Escalating too early: Everything becomes an Inquiry, which defeats the purpose and overloads your team.
No context pack for humans: Staff re-ask basics, customers repeat themselves, and response time increases.
Unclear ownership: Inquiries sit unassigned, so customers feel ignored even though you “escalated.”
Not updating training inputs: The same questions keep escalating forever, so the system never improves.

FAQ

How does the workflow prevent the Agent from guessing?

Enable the Inquiry behavior so when the Agent can’t answer confidently, it escalates and logs the case—rather than improvising.

Can a human take over the exact WhatsApp conversation?

Yes. Clarivo supports an “open in WhatsApp” style human handoff, so a team member can jump into the same thread and continue.

Does this work for multilingual customers and voice notes?

Yes. Clarivo automatically replies in the customer’s language and can understand WhatsApp voice notes, showing transcription in the dashboard.

Do I need to build on the WhatsApp API to do this?

For the core workflow, Clarivo connects via QR code (linked-devices style) and uses your existing WhatsApp Business setup.

What should the Agent do while waiting for a human?

Keep it simple: acknowledge escalation, set expectations, and (only when appropriate) gather the minimum details needed for routing.