The best lead qualifier does not feel like an exam. It resembles a short conversation with somebody who understands the business and asks only questions whose answers change the next step. Quality falls as soon as the form confuses curiosity with necessary data collection
An interview with the enquiry
What should be known after qualification?
Not whether a lead is “good” or “bad”. The system should know which route fits: a direct specialist call, a prepared appointment, a standard offer, a partner referral or a polite decline
Which question comes first?
One that provides orientation and is easy to answer. For a software project, that may be the affected process. For a technical service, the location and type of problem. Budget or telephone number as the first hurdle often looks like a filter designed for the supplier
How many questions are right?
Enough to distinguish the routes safely. The number alone is not useful. Six precise questions can be harder than ten simple ones when nobody understands why they are being asked
Questions must earn a decision
Write a small card for every field:
- Which decision uses this answer?
- Can the prospect know it at this point?
- Which answer options are missing from real cases?
- Is free text necessary or merely convenient for the author?
- How long may the information be retained?
If no decision can be named, remove the field. This eliminates a surprising number of questions about headcount, position and “How did you hear about us?” that are collected but never used
Good branching explains itself
After “existing system”, a question about the current software can appear. It would be meaningless after “new build”. Branches should follow previous answers without leading users into a dead end. Visible progress and a way to correct answers keep the conversation understandable
Going back should not silently erase data. If a change invalidates later answers, the system explains which sections need review
Uncertainty is a valid answer
In complex B2B work, a prospect may not yet know the budget, integrations or schedule. Without an “I do not know yet” option, the form forces invented precision. The uncertainty is useful information in itself. It indicates that a discovery call or design phase should come before a proposal
A free field can also be valuable when it asks something specific: “What must work better in daily operations six months from now?” This produces more substance than “Additional message”
No black box for sales
A rule set can suggest a route from the answers. The team must be able to see why. “Direct call because budget exists, decision maker is identified and delivery is needed within three months” can be reviewed. An unexplained AI score of 82 cannot
Language models can summarise free text or suggest themes. They should not trigger a final rejection unless training, error rate and potential bias have been examined. The original text remains visible to accountable staff, and sensitive information is not sent to external models without review
The polite alternative to a dead end
Enquiries that do not fit deserve a concrete next option: documentation, a partner, a smaller package or a date for reconsideration. That is more honest than allowing every enquiry to book a meeting that will later be cancelled
The qualifier learns from decisions
After several weeks, compare the suggested route with the actual outcome. Which highly ranked leads never became concrete? Which declined enquiries developed into suitable work? The result changes questions and rules. It is not a tool for monitoring salespeople by close rate
Lead scoring for B2B enquiries explains calibration and error rates for a quantitative next step. If the result should open a calendar directly, see appointment booking with qualifying questions. The qualifier itself remains a good conversation expressed in software, not a bouncer