Lead qualification calls

A new inquiry is not yet a sales opportunity. The caller may be outside the service area, need something the business does not offer, have a timeline the team cannot meet, or simply want basic information. When every representative asks different questions, the resulting records are hard to compare and good prospects can wait behind incomplete ones.

An AI phone agent can run a consistent first conversation. It asks the approved questions, confirms important details, records the answers, and follows explicit routing rules. It does not need to decide whether a person is a “good lead” in the abstract. The business defines lead qualification and which conditions warrant a transfer, booking option, follow-up, or polite close.

This function is useful across solar installers, professional services, home services, and other teams, but the page is about the call workflow rather than any one industry's intake script.

A concrete call walkthrough

Suppose a caller asks for a project estimate.

Agent: I can collect a few details and connect you with the right team. What are you hoping to get done?

Caller: I want an estimate for work at a commercial property.

Agent: What is the property ZIP code, and when would you like the work to begin?

The agent follows the prompt's question order, but it can keep the exchange conversational. It confirms the caller's name, contact information, location, request type, timing, and any other fields the business has approved. A connected knowledge base can supply factual answers about service areas or standard offerings; the agent should escalate rather than improvise when the approved material does not answer the question.

Assume the ZIP code is covered, the requested service is offered, and the timeline meets the team's rule. The agent creates or updates the lead through the connected integration and attempts the configured handoff.

Agent: Your project matches the commercial team. I can try to connect you now. You may wait briefly while I explain the request to them.

On an eligible production number, attended warm transfer places the caller on hold while the agent privately briefs the recipient. If the team declines, does not answer, or reaches voicemail, the agent returns to the caller with the outcome and follows the fallback in the prompt. If the lead falls outside the defined criteria, it uses approved language and records the reason rather than inventing an exception.

How to set it up on ThunderPhone

Write the prompt around objective fields and outcomes. A useful outline includes:

  • the disclosure and opening question;
  • required contact and request fields;
  • qualification criteria such as service type, geography, timing, or account status;
  • questions the agent may answer from approved material;
  • conditions for immediate transfer, later follow-up, or an out-of-scope close;
  • statements the agent must not make, including promises about price or availability; and
  • the final summary it should confirm with the caller.

Attach only the relevant knowledge-base documents to the agent. ThunderPhone supports text, Markdown, CSV, PDF, and DOCX files, content search, per-agent document scoping, and mid-call knowledge search. Uploads do not apply automatically; select the approved documents for this agent.

Use HubSpot or Salesforce to store the lead in an existing sales process. Google Sheets can be appropriate for a simple, controlled intake table. Slack can notify the right team of an urgent qualified inquiry. For a custom system, use REST, webhooks, or a remote MCP server. ThunderPhone can connect to remote MCP servers over streamable HTTP, discover their available tools, and attach selected servers to an agent. Expose narrow operations such as finding a record or submitting approved intake fields rather than unrestricted system access.

Choose an eligible production number if the workflow needs warm transfer. Cold transfer is also supported; demo numbers and browser calls are limited to cold transfer. Test ambiguous answers, unsupported locations, duplicate contacts, callers who refuse a required field, and integration failures before opening the number to live traffic. Reusable scenarios and regression gates help show whether a prompt change still follows the rules.

What to measure

Measure each stage separately:

  • calls with all required qualification fields completed;
  • inquiries meeting each explicit criterion;
  • transfers attempted, accepted, and returned to the caller;
  • records successfully created or updated;
  • duplicate or incorrectly matched lead records;
  • callers routed to follow-up, booking, or an out-of-scope outcome; and
  • later sales outcomes by initial call disposition.

Avoid using call length as a stand-in for quality. A short call can be complete, while a long one can still be missing the decision fields. Compare qualification results with phone intake forms and client intake when choosing how much detail belongs in the first call.

Honest limits

The agent should not infer budget, intent, authority, or fit from tone. It should not make unapproved pricing promises, negotiate terms, or mark a lead qualified when a required answer is missing. Complex buyers, strategic accounts, hostile callers, unusual requests, and anyone asking for advice outside the approved material need a human.

Qualification is also not the same as speed to lead. Fast contact matters, but consistent evidence matters too. Keep the rules inspectable, review false positives and false negatives, and let sales staff own judgment that cannot be reduced to explicit criteria. For the adjacent callback workflow, see speed-to-lead callbacks.

Configure a ThunderPhone agent for lead qualification