Hours, directions, and info line
Location questions look easy until a business has holiday schedules, multiple entrances, appointment-only departments, parking restrictions, or sites with similar names. A caller may ask, “Are you open?” when the real question is whether a particular desk is open, whether they can arrive late, or where an accessible entrance is located. A generic recording rarely handles that context, while sending every question to the front desk interrupts higher-value work.
An hours, directions, and info line gives callers a conversational way to reach approved location information. The AI phone agent identifies the location and purpose of the visit, searches the business’s source material, reads back the relevant details, and routes questions that require judgment. It can work as a focused line or alongside 24/7 phone answering.
A concrete information call
Imagine a service center with a public entrance, a separate pickup counter, and different Saturday hours.
Agent: I can help with hours, directions, and arrival information. Which location are you visiting?
Caller: The Northside location. I need to pick up an item on Saturday.
Agent: The approved listing shows that the Northside pickup counter is open Saturday from 9 a.m. to noon. The main service desk is closed. Are you driving or arriving on foot?
Caller: Driving, and I use an accessible parking space.
Agent: The arrival instructions say to use the east lot from Pine Street. Accessible spaces are beside the pickup entrance. I can repeat the street address and entrance instructions slowly.
Caller: I may not get there until 12:15. Will someone wait?
Agent: I cannot promise service after the published closing time. I can connect you with the location to ask about an exception.
The call stays useful without turning published information into an operational guarantee. The agent distinguishes the pickup counter from the main desk, gives only the stored arrival instructions, and escalates the request that requires a person at the location.
How to configure it on ThunderPhone
Create a prompt that resolves ambiguity before giving an answer. It should instruct the agent to:
- Ask which location, department, and visit type the caller means.
- Search the approved source rather than rely on a general statement about the business.
- State the date when giving holiday or temporary hours.
- Read addresses, suite numbers, landmarks, and entrance names in short segments and offer to repeat them.
- Avoid guaranteeing travel time, parking availability, walk-in service, or an exception to posted hours.
- Transfer or record a follow-up when the source does not cover the caller’s situation.
Store the location directory in ThunderPhone’s document library. Text or Markdown works for narrative arrival instructions; CSV can keep repeated fields such as location name, address, department, weekday hours, and accessibility notes consistent. The library also accepts PDF and DOCX. Select the relevant documents for this agent, because uploads do not apply to agents automatically. The built-in knowledge-base search can retrieve the matching location during the call.
Use a production number for the public line. An inbound demo U.S. number can support testing, but it is not intended for production. If hours already live in a business system, an approved REST tool can expose the current record to the agent. Use a narrow response schema and define what happens when the system and the written source disagree. A Google Sheets connection may suit a reviewed location table; REST or webhooks can connect a custom source. Do not make the agent guess which source is newer.
For exceptions, Slack can alert an assigned team or a transfer can route the caller to the location. Attended warm transfer requires an eligible production number that can place an outbound SIP leg. Otherwise use cold transfer or capture the caller’s question for follow-up. The prompt must give a clear fallback when a location does not answer.
Test similar location names, a holiday, a department with different hours, an address spelling request, accessibility questions, severe-weather questions not covered by the source, and a caller seeking an after-hours exception. Graded scenarios should fail if the agent invents live traffic, parking occupancy, closure, or travel-time information.
What to measure
Track the reasons people call and how reliably the line resolves them:
- calls completed with approved information;
- transfers or follow-ups by location and reason;
- searches that found no matching source;
- repeat requests to spell or clarify an address;
- calls caused by conflicting or outdated listings; and
- call-review findings where the answer exceeded the source.
The unanswered-question list is also a content maintenance queue. If many callers ask about the same entrance or department, improve the location record. If they need broader policy explanations, route them to an FAQ answering flow rather than overloading the directions prompt.
Honest limits
This line cannot observe a locked door, a full parking lot, an unreported closure, or current road conditions unless the business supplies that information through an approved source. It should not promise that a late arrival will be admitted or that a walk-in will be served. Emergency reports, safety hazards, lost people, and accessibility problems requiring immediate accommodation need a human. Configure a clear urgent-call routing path for those cases.
For a caller who needs to reserve a time rather than learn when a location is open, use appointment scheduling with its own availability and confirmation workflow.
Create your ThunderPhone account to build and test an hours, directions, and info line.