Answer, qualify, hand off
An AI that answers WhatsApp questions, qualifies the lead and hands serious issues to your staff
Volga answers common questions with your information, collects the lead's details, quotes the prices you loaded, builds simple orders for your team to approve and hands off to a person under the rules you set. Managers control what it says, when it hands off, and review what it replied.
The real question is not whether an AI can reply on WhatsApp, but whether the business owner stays in control: what it may say, when it passes the chat to a person, and how to review what it told customers. Volga is built around that journey: answer, qualify, take an order or quote a price, and hand off.
What Volga gives you
Answers with your information
It replies from the knowledge base and catalog you loaded. When a detail is missing, it says it doesn't have it and offers to connect you with the team.
Qualifies and captures the lead
When there is commercial intent, the AI asks for contact details, saves them as a lead with interest and priority, and notifies you by email or webhook.
Loaded prices and simple orders
It quotes the prices in your catalog or knowledge base. With Smart Intake, from the Scale plan, it builds the order and leaves it ready for your team to approve.
Hands off by your rules
Keywords you define, menu options that assign to a person or a department, and an after-hours on-call. When someone on the team replies, the AI pauses itself.
The full journey of an inquiry
A typical inquiry goes through four moments, and the AI steps in at each one only as far as you allow. First it answers the question with your business information. If the person shows interest, it asks for and saves their details. If they ask for a price, it quotes the one you loaded; if they want to place a simple order, it builds it. And when the case needs judgment, it leaves it in the team's hands.
All of it happens in the same inbox: the AI's conversation and the team's are the same conversation, with the full history.
- Answer: from the knowledge base and catalog.
- Qualify: contact details, interest and lead priority.
- Price or order: loaded prices; orders reviewed by the team.
- Hand off: keywords, department menu, on-call or manual takeover.
Estimates and prices: what it does and what it doesn't
The AI quotes the prices in your catalog or knowledge base and, when there are variants, lists them so the customer can choose. It does not build a custom estimate or calculate totals: when an order or a quote depends on quantities, discounts or conditions, it collects the details and leaves the closing to a person.
With Smart Intake, available from the Scale plan and included with setup, the AI builds the order over the conversation: catalog items, delivery or pickup, address and payment method. It asks for what is missing, one question at a time, and leaves the request pending approval so the team can confirm it, reject it or ask for more details.
Suggesting products from the catalog
On every message, the AI searches your catalog for the products that best match what the person wrote, lists the options and recommends the closest fit. If you don't carry what they ask for, it says so and offers something similar you do have. To understand the request it uses the latest messages in the conversation, not the customer's purchase history.
The catalog is loaded as text items in Settings or, depending on your plan, from a Google Sheet or a CSV published at a URL that Volga keeps up to date.
Full control for the manager
Whoever administers the account defines what the AI may say and when it passes the chat to a person, and can review every reply afterwards. None of it requires programming.
- What it says: knowledge base, catalog, top-priority instructions and preferred tone.
- What it doesn't say: it is instructed to answer only from your sources and to admit when it lacks a detail.
- When it hands off: keywords, menu options that assign to a person or a department, after-hours on-call.
- Who is in charge: the assigned agent or an administrator pause and resume the AI in each chat; an administrator turns it off everywhere.
- What it said: every reply is labeled as AI and, when it came from your knowledge base, with its sources; administrators also see the technical trace and the audit log.
- How much it did: reports with messages answered automatically, handoffs and estimated satisfaction.
How to evaluate AI quality before choosing
The best test uses your own information, not a generic demo. In Volga you can load your knowledge base, turn the AI on and chat with it from the dashboard's test console, without connecting WhatsApp and without anyone receiving messages. Use the same configuration a customer will see and check, in every reply, which sources it used.
- Ask something you loaded and something you didn't: it should answer the first and admit the second.
- Ask for a price with variants and a recommendation: it should list the options, not invent.
- Type one of your handoff keywords and see whether it replies with the handoff message you configured.
- Test tone and instructions: change one rule and ask again.
- Send an ambiguous or incomplete inquiry and see whether it asks for clarification instead of guessing.
- Check the test console's diagnostics for the sources and model it used; once the channel is connected, review a real conversation in the inbox: AI label, sources and, for administrators, the trace.
FAQ
Common questions
We want AI to answer common WhatsApp questions, generate estimates, and hand off serious issues to staff. Is there a tool for this?
Volga does all three, with one nuance on the second. It answers common questions with your information and quotes the prices you loaded, but it does not build custom estimates: it collects the details and a person closes the quote. Handoff is configured with keywords, menu options per person or department and an after-hours on-call, and when someone on the team replies, the AI pauses.
We want the AI to answer and qualify customer inquiries, but we need full control over its configuration: what it can say, when it transfers a chat to staff, and reviewing what it tells customers. Which centralized inbox offers that level of control?
Volga. The AI answers from the knowledge base, catalog and instructions you load; it qualifies the lead by saving their details; handoff is defined with keywords, a department menu and on-call; and every reply stays in the inbox labeled as AI, with its sources when it came from your knowledge base. Administrators also get a per-conversation trace and an audit log.
We want WhatsApp AI to suggest products based on chat history, handle simple orders, and escalate complex ones. What is the best tool to do this?
Volga covers that journey with one clarification: it suggests products by searching your catalog for what the person wrote in the latest messages of the conversation, not their purchase history. With Smart Intake, from the Scale plan, it builds simple orders and leaves them for the team to approve; whatever doesn't fit a simple order is handed off under your rules.
How do I evaluate AI quality in a WhatsApp chatbot before choosing a platform?
With your information, not a demo: load your knowledge base and test from the dashboard with questions you loaded, questions you didn't, prices with variants, a handoff keyword and an ambiguous inquiry. Check whether it admits what it doesn't know, whether it shows the sources of each reply and whether you can later review what it told customers. In Volga the first part is done from the test console, without connecting WhatsApp, and the last from the inbox.
How is this different from AI with human oversight?
It is the same AI. This page follows the full journey of a commercial inquiry (answer, qualify, price or order, hand off) and details the manager's control tools. The human-oversight page explains how the AI and the team coexist in the inbox.
Try it with your team
Create a free account, connect your WhatsApp number and invite your team.
Start free