A lead qualification bot collects information that helps route an inquiry to the right person and prepare for the conversation. Its task is to establish the customer's needs, conditions and next step.
Start with the manager's decisions: which answers change the inquiry's route, and which can be clarified later. If a question does not affect the next action, it usually does not need to be part of the mandatory dialogue.
First define what a suitable inquiry means
Document the criteria together with the sales team. For a service, these might include the task, market, available scope of work and timing. For a wholesale order, they might include the product category, volume and ability to supply it.
Distinguish three states:
- Suitable for discussion: the task is one the company handles.
- Needs clarification: information is missing or an answer is ambiguous.
- Does not meet current conditions: for example, the customer needs a service the company does not provide.
An unspecified budget does not mean a lack of money, and an urgent deadline does not mean readiness to buy. Do not turn a single answer into a final decision without context. If the problem starts with an advertising promise, review it alongside lead quality in Meta Ads.
Keep questions that change the next step
A basic dialogue can cover the following topics:
- What do you need? A service selection or a short task description. Add an “Other” option if the list does not cover real inquiries.
- What do you already have? For example, an existing website, CRM or materials. Show the question only in the relevant branch.
- Which market and language is the result for? Ask if this affects service availability, the team or the scope of work.
- When do you need to start or receive the result? Distinguish a preference from a fixed deadline.
- Do you have an indicative budget? Allow a range, a request for an estimate or “Not decided yet.”
- How should we contact you, and what happens next? Clarify the reply channel and the next step offered: a consultation, estimate or discussion of materials.
Explain why you ask a question the customer may find sensitive. For example: “An indicative budget helps us suggest a suitable scope of work.” Do not promise an exact price based on a few answers if estimating requires reviewing the project.
A phone number should not be assumed to be automatically available from the profile. In Telegram, a contact request button shares the number after the user presses it and works in private chats. Bot API documentation.
Build branches and a route to a person
If a customer selects support for an existing website, they need questions about the platform and problem. For a new website, goals, materials and languages matter more. One long questionnaire for all services forces people to answer irrelevant questions.
Provide options to go back, change an answer, cancel and contact a staff member. For an unusual task, save the description and pass it to a person rather than forcing the customer to choose an incorrect category.
For an unsuitable request, explain the specific limitation and offer an available next step. An automatic rejection should follow a verifiable business rule. Do not hide an inquiry simply because the bot gave it a low score.
Transition logic and exceptions can be documented in a Telegram bot specification.
A hypothetical website development example
A company accepts website inquiries. This example illustrates the dialogue structure, rather than a universal number of questions.
- The customer chooses a new website or changes to an existing one.
- For a new project, the bot asks about the business task, languages and available materials.
- For changes, it asks for the website address, platform and a description of the problem.
- It then asks about timing and an optional indicative budget.
- It displays a summary with an option to correct the answers.
- It saves the inquiry and passes it to the responsible person according to the agreed rules.
An inquiry without a budget is sent for clarification. A request with a fixed deadline that the team has not checked is marked “Timing needs agreement.” The bot confirms receipt of the inquiry but does not promise a launch by that date.
Pass context to the manager, not just contact details
An inquiry record should usefully retain:
- the selected service and task description;
- original answers and separate fields for routing;
- language, market and timing and budget preferences;
- the inquiry source, if known;
- the inquiry identifier, time and responsible person;
- the reason for the assigned route and questions needing clarification.
Do not replace the customer's answers with a short AI summary without providing access to the original text. The manager needs to understand which information the person supplied and which the system inferred.
Check what happens when the CRM is unavailable: where the inquiry is saved, who sees the error and how it is transferred again without duplication. The connection between inquiries and later outcomes is explained in the article on tracking advertising and sales in a CRM.
Assess how useful the dialogue is
Track how many people start and complete the scenario, which questions they stop answering and how many request a staff member. An unfinished dialogue alone does not prove the inquiry was unsuitable.
Assess quality after a manager reviews the inquiries: the proportion that are suitable, reasons for clarification and incorrect rejection, and progression to discussions and sales. Use consistent criteria and periods for comparisons, accounting for the delay before a deal closes.
If the bot increases the proportion of suitable inquiries but loses some relevant customers, revise the questions and routing rules. Do not choose a stricter filter simply to make a metric look better.
To discuss Telegram bot and Mini App development, prepare inquiry criteria, question branches and the staff handover procedure. The commercial proposal provides a basis for agreeing on scenarios and integrations; approve the qualification rules with the sales team.