The problem with chatbots that ask the wrong questions
Most chatbots follow a rigid script: the same questions, in the same order, for everyone. A buyer ready to sign and a casual browser get the exact same journey. Result: the former gets impatient, the latter drops off, and conversion rates stall.
Structured qualification changes that. According to Téléaction, a well-built qualification method can take a B2B pipeline's conversion rate from 8% to 31%.
The 3-phase system: Identification → Qualification → Conversion
This is the pattern we build into every H'appi bot. It was born from the INnatural project (natural cosmetics): our client's chatbot was asking product questions before even knowing whether the visitor wanted to buy, resell, or just browse.
Phase 1 — Identification: who are you actually talking to?
Before qualifying anything, the bot determines within 1-2 exchanges:
- Visitor profile (individual, business, reseller, potential partner)
- General intent (purchase, support, information, recruitment)
- Perceived urgency in the message
This first pass avoids the classic mistake: asking sales-qualification questions to someone who just wanted to know the opening hours.
Phase 2 — Qualification: the right question, at the right time
The B2B reference method is still BANT (Budget, Authority, Need, Timeline). It's solid, but applied literally in a chat, it feels like an interrogation — four closed questions in a row drives people away more than it converts them.
Our approach: qualify progressively, building on previous answers rather than a fixed form. If the visitor already mentions a budget or timeline in their first message, the bot doesn't ask again — it moves straight to the next step.
Phase 3 — Conversion: offer the action that actually makes sense
A hot lead and a cold lead should never get the same button. Depending on the qualification level reached, the bot offers:
- Hot lead, clear budget and timeline: direct meeting booking with a sales rep
- Warm lead, identified need but unclear timeline: targeted documentation + email capture for follow-up
- Casual browser: answer their question, without forcing conversion
What we learned with INnatural
After deploying this 3-phase system on the INnatural widget, the visit-to-qualified-lead conversion rate improved by 34%. The most impactful change wasn't a specific question — it was the order: identify before qualifying, qualify before converting.
This pattern is now the default skeleton of every H'appi bot, adapted to the vocabulary and qualification criteria specific to each industry.