Zanus AI After Hours for Heating Firms: Test the Human Backup

Why After-Hours Answering Matters for HVAC Operations

Heating emergencies do not respect business hours. A furnace failure in December or an air-conditioning outage in July can happen at 11 PM, on a Sunday or during a holiday. The homeowner who cannot reach anyone often calls a competitor instead. According to Zanus AI's own material, roughly a third or more of inbound calls to home service companies arrive outside office hours, and these are the highest-intent ones: no heat, no water, an alarm going off. For a heating firm, the after-hours period is not a gap to be managed with voicemail; it is a revenue channel that either captures the customer or loses them.

A traditional voicemail system leaves a message and a callback request. The customer waits, uncertain whether anyone is listening. An answering service picks up during staffed hours but requires a person in a call center somewhere, scaling with the number of incoming calls. AI Phone offers a different model: an AI receptionist answers on the first ring and either resolves the call or announces the caller to an on-call technician or manager and waits for acceptance before connecting them. For a heating firm, this means the first filter is instant, and the human decision about whether to accept or defer the call remains with your team.

The Core Difference: Announced, Accepted Transfer

The distinction between a traditional transfer and what Zanus AI calls an "announced, accepted transfer" is the foundation of your test plan. Here is how it works in practice:

When a call comes in after hours and the AI cannot handle it alone, the system announces the caller and the reason to the on-call person, who then presses 1 to accept the call. If the person does not accept within the timeout window, or if a mobile voicemail is about to pick up and swallow the call, Zanus AI pulls the call back to the AI and offers the caller a message option instead. This prevents the silent failure where a call rings your technician's phone, goes to voicemail, and the homeowner never hears back.

Compare this to a simple transfer: the AI rings your phone, the call connects whether or not you accept, and if you are asleep or unable to answer, the caller hears your personal voicemail and may not realize the company is trying to reach them. The announced model forces an explicit decision: your on-call person knows a call is coming and must choose to take it. If they do not, the system handles the fallback.

Pre-Test Setup: Ownership and Department Policy

Before you run any test, establish clear ownership and decision authority. For a heating firm testing AI Phone, assign one person to own the on-call setup and another to own the call log review. This prevents confusion about whether the test ran correctly or whether the system failed.

Define your department policy in writing. For example:

The second and third rows are what you test. The fourth row is for gas-smell, flooding or electrical-hazard calls that require immediate escalation. Zanus AI can be configured to read an approved safety script word for word—such as "leave the house and call the gas utility"—and then transfer to your on-call technician. You never write this script yourself; your firm's safety or legal responsibility defines it, and you approve it before Zanus AI uses it.

Building the Test Scenario: The Unavailable On-Call Technician

The most revealing test is the one where your on-call person is genuinely unavailable or slow to respond. This is not a failure of the system; it is the system working as designed. The test is whether the fallback succeeds.

Here is an illustrative scenario: It is 10 PM on a Thursday. The dispatch department is closed. A homeowner calls to report that their furnace is not heating. The AI should recognize this as an after-hours call and, depending on your policy, either attempt to book a callback visit or escalate to the on-call technician. If your on-call tech is asleep, traveling or offline, the test is whether Zanus AI:

1. Announces the call to the on-call tech's mobile with the caller's name and the reason for the call.
2. Waits for the tech to press 1 to accept, within a timeout you specify (typically 20–30 seconds).
3. If no acceptance, pulls the call back and offers the homeowner a message option, recording the callback number read back and confirmed digit by digit.
4. Delivers the message to the tech's own inbox with the homeowner's context and callback number.
5. Logs the call outcome—answered, accepted transfer, message created, staff response pending—so you can review it the next morning.

The last point is critical. Zanus AI transcribes every call and writes a summary into the customer's record. That record becomes your morning briefing and your audit trail for what happened overnight.

Test Case 1: Answered and Accepted Transfer

Set up a test call during the hours when your on-call person is genuinely available and can respond quickly. The purpose is to confirm that the announcement works and that your team member can accept the call without confusion.

Illustrative record of a successful transfer:

In this scenario, the on-call tech heard the announcement—"Sarah Chen is calling about no heat and a furnace that won't start"—and accepted by pressing 1. The call connected and the tech was able to book a visit for the next morning and send a confirmation. This is the happy path. Your test confirms that the announcement is clear, the acceptance mechanism works on a mobile phone, and the handoff does not drop the call.

Test Case 2: Message Created When No One Accepts

The harder test is when no one is available. Schedule this test when your on-call person is intentionally not reachable: asleep, in a meeting, or with Do Not Disturb enabled on their phone.

Illustrative record when acceptance fails:

The acceptance criterion here is: Did the AI pull the call back before it went to the tech's voicemail? Did it offer the homeowner a message option? Was the callback number read back and confirmed? Is the message now in the on-call tech's inbox, timestamped and linked to the customer record?

If all five are yes, the system passed the test. The homeowner has a recorded message, you have a digital trace of what happened, and the on-call tech will see it in their inbox when they wake up or finish their meeting.

Test Case 3: Holiday Rota and Stale Schedule

Heating firms often have shifted coverage during holidays. Christmas, New Year and summer vacation weeks may have different on-call rosters than normal weeks. This test confirms that Zanus AI uses your updated schedule and does not route to someone who is actually off.

Before the test, update your on-call directory to reflect the holiday rota. For example, on December 24–26, Marcus is off, and Jessica takes over on-call duties. Zanus AI stores an operator's directory with per-day working hours and vacation dates, and you must keep this current. If you do not update it, a call on December 24 might still ring Marcus, who is asleep at home and will not accept because he is not supposed to be on duty.

Test action: Schedule a test call during a holiday with the updated rota active. Call your main number at 11 PM on December 24 and simulate an emergency request. The AI should either attempt to reach Jessica (the correct on-call person) or, if Jessica is also off, play the message you wrote for that closure day and offer a message option.

Acceptance criteria:

• The announcement goes to the correct on-call person for that day.
• If all on-call staff are off, the AI plays your pre-written holiday message, such as "We are closed for Christmas. In a life-threatening emergency, call 911. Otherwise, we will return your call on December 27."
• The homeowner is offered a message option, not transferred to someone unavailable.

Test Case 4: Returning Caller and Callback Failure

A second test is what happens when a caller phones back after leaving a message and your on-call person still has not replied. This test reveals whether Zanus AI recognizes a returning caller and whether it escalates or repeats the same flow.

Illustrative scenario: On Friday at 11:30 PM, a homeowner calls about frozen pipes. The on-call tech does not accept. A message is recorded. On Saturday at 8 AM, the homeowner calls back because they still have not heard from anyone. Zanus AI should recognize the number as a known caller and greet them by name. Your policy should define what happens: does the AI offer them a message again, or does it escalate automatically because they have already called twice?

Include a returning caller in the controlled routing trial. Ask the supplier to demonstrate the available matching and routing behaviour in the proposed configuration rather than assuming the call follows a different route.

Test action: Call once, leave a message. Wait 30 minutes. Call again from the same number. Observe whether Zanus AI recognizes you and handles the second call according to your escalation policy.

Test Case 5: Weekend Overflow and Unexpected Volume

Demonstrate simultaneous calls within the proposed deployment’s actual capacity. Confirm what callers hear when the arrangement reaches a limit and how the office recovers unresolved requests.

With AI Booking enabled, Zanus AI integrates with your calendar and offers only the free slots to callers. If your weekend schedule shows that your only available technician can handle one emergency on Saturday morning at 9 AM, and two homeowners both call, the AI should book the first one for 9 AM and offer the second one a later time or a callback when another slot opens.

Test action: Simulate three back-to-back calls on a Saturday with limited dispatch availability. Confirm that Zanus AI does not double-book and that homeowners receive realistic arrival windows or callback estimates.

Test Case 6: Distinguishing Emergency From Routine Call

Not all after-hours calls are emergencies. A homeowner might call to schedule a maintenance visit or to ask a question about their bill. Zanus AI needs to triage these correctly. A gas-smell call is an emergency and goes immediately to the on-call tech. A routine appointment request might be handled entirely by the AI or taken as a message for follow-up during business hours.

For heating firms, Zanus AI can be configured with an approved safety script for true emergencies: gas smell, no heat when freezing outside, visible electrical sparking, or water leaks. When a caller mentions any of these, the AI reads your safety script word for word—"If you smell gas, leave the house immediately and call your utility company's emergency line"—and then transfers to the on-call tech. For all other calls, the AI follows your routine after-hours policy.

Test action: Place one call reporting a gas smell and one call requesting a routine appointment. Confirm that the gas-smell call escalates immediately and the appointment call is either booked or logged as a message for business hours.

Setting Up the On-Call Directory and Approval Workflow

Zanus AI requires you to build and maintain an operator directory. This is not a one-time setup; it is an ongoing responsibility. Each person must have:

• Extension and mobile number.
• Department assignment (e.g., dispatch, technical support).
• Languages spoken.
• Working hours for each day of the week.
• Vacation dates and temporary expiry dates for seasonal workers.
• Nickname or preferred greeting, if different from their legal name.

Place controlled test calls to each intended mobile destination and verify the result. Check both an available recipient and the configured fallback when that person cannot accept.

The approval workflow is: You talk, Lisa (the AI assistant that sets up Zanus AI) confirms the change, and you say yes before it is applied. This prevents accidental misrouting. For example, if you tell Lisa "Marcus is on vacation next week," she will say "Marcus's hours are set to zero next week; is that correct?" You confirm, and then the change is live.

Acceptance Criteria and Change Decisions

Before you run your test, write down the specific outcomes that will make you buy or not buy Zanus AI Phone. Here are examples of acceptance criteria for a heating firm:

• Every after-hours call is answered on the first ring, with no queue or hold.
• Calls that require a transfer are announced to the on-call person with the caller's name and reason, and the person must press 1 to accept.
• If a person does not accept within 30 seconds, the call is pulled back to the AI and a message option is offered.
• Every message is transcribed and delivered to that person's inbox with the callback number read back and confirmed.
• Returning callers are recognized and handled according to policy.
• Emergency calls are escalated immediately; routine calls are logged for business hours.
• The call log shows a clear outcome for every call: answered, accepted transfer, message created, booked appointment or escalated.
• Setup takes less than one hour and does not require IT support beyond forwarding your phone line.

If Zanus AI meets all of these in your test, the purchase decision is straightforward. If it fails one or more, quantify the impact: Is it a deal-breaker, or is it a configuration issue you can resolve with support? For example, if a call is not pulled back fast enough and a voicemail answers instead, that is a technical failure. If your on-call tech's number is wrong in the directory, that is a data error you need to fix.

Conducting the Seven-Day Free Trial

Zanus AI offers a seven-day free trial with no credit card required. The trial gives you a real number and real call-answering capability on your business. The key is to run your test calls during this week and to review the call log when the trial ends.

Your trial plan:

• Day 1: Set up your business profile, department policy and on-call directory with Lisa. Confirm that your main phone line is forwarded to the Zanus AI number.
• Days 2–4: Run test cases 1 and 2 during hours when you can control the on-call availability. Document every call outcome.
• Day 5: Test case 3 (holiday or shifted rota).
• Day 6: Test cases 4, 5 and 6 with real-world scenarios or simulations.
• Day 7: Review the call log, export the data and summarize the outcomes against your acceptance criteria.

At the end of the trial, Zanus AI sends you a report of every call: who called, when, the outcome and whether the AI handled it alone or passed it to a person. Use this report to decide whether the system meets your needs.

Common Failure Modes and Troubleshooting

Several issues may arise during your test. Here is how to distinguish a system failure from a configuration problem:

Voicemail is answering instead of Zanus AI. This usually means your phone line is not forwarded correctly, or the forwarding rule is not active. Confirm with your phone provider that calls are routing to the Zanus AI number you provided. Zanus AI can help you test the routing.

The on-call person is not hearing an announcement; the call is connecting silently. This usually means the transfer settings are not configured to announce. In your policy, confirm that the department is set to "announced transfer" or "operator first" and not "AI only." Ask Lisa to review your transfer settings.

A message is recorded but not delivered to the person's inbox. Check that the person's email is correct in the directory and that they have logged into Zanus AI at least once. Messages are delivered to their "My phone" inbox in the Zanus AI interface, not to their email, unless you have enabled email notifications.

The AI is booking appointments at times when no one is available. This means your calendar is not synced or is out of date. Confirm that your dispatch calendar is connected to Zanus AI Booking and that availability is accurate for the dates you are testing.

A returning caller is not recognized. Zanus AI matches by phone number. If the caller is using a different phone or the number is not in your database, they are treated as a new caller. This is correct behavior; you cannot greet someone by name if you do not know their number.

Pricing and Implementation Scope

Zanus AI Phone is priced per company, not per seat. You can add all your technicians, dispatchers and managers to the on-call directory without paying extra per person. The pricing includes transcription, message storage and call analytics. If you add AI Booking, you pay for that module separately. See the pricing page for current rates.

Implementation scope for a heating firm typically includes: one phone number (your existing line), one department (dispatch), one to three on-call people and a basic safety script if you handle emergency calls. This is minimal configuration and should take less than one hour to set up with Lisa.

If you later add multiple locations, multilingual support or AI Quoting (for flat-rate estimates), those are additional modules that integrate with your phone line. You do not need to change anything on the phone side; you add modules through Zanus AI's interface.

Key Differentiators for HVAC and Heating Services

Zanus AI's feature set is built with home services in mind. Here are the differentiators that matter most for heating firms:

Announced transfer with acceptance means your on-call tech hears who is calling and why before the call connects. This is not true of all AI answering services, which may route calls blindly or use a queue model that does not distinguish between callers.

Safety script integration allows Zanus AI to read your approved gas-smell or electrical-emergency script without embellishment, and then escalate to the on-call technician. You write or approve the script; Zanus AI does not compose emergency guidance.

Booking from your calendar ensures that the AI does not overcommit your dispatch capacity. A homeowner cannot book a time slot that does not exist in your schedule.

Message transcription and delivery to individual inboxes means your on-call person sees a written record of the call, not just a voicemail they have to listen to. This is faster and creates a permanent record for compliance and follow-up.

Frequently Asked Questions

Can Zanus AI dispatch a technician automatically, or does a person still need to assign the call? Zanus AI books the appointment into your calendar and creates a service record, but your dispatcher still assigns the truck and technician. Zanus AI integrates with call analysis and your CRM, so the appointment appears in your normal workflow. It does not replace your dispatch board; it feeds it.

What happens if an on-call technician accepts a call but their mobile signal drops? The call ends. Zanus AI does not reconnect automatically. Your technician can call back the homeowner's number from their own phone, or if they need to restart, they can call your main line and the AI will recognize them as an internal caller.

Does Zanus AI work with my existing phone system? Yes. Zanus AI sits in front of your phone line. You forward your business number to Zanus AI's number, and the AI answers. Your internal phones, extensions and transfers stay the same. If you later port your number to Zanus AI, they handle the technical work with your carrier.

Can I review calls after hours to make sure everything was handled correctly? Yes. Every call is logged with its outcome, transcript and summary. You can review it in real time on the call log, or wait for the morning briefing, which reads you a summary of overnight calls.

Is there a limit to how many calls Zanus AI can answer at once? No. Zanus AI is cloud-based and can answer multiple simultaneous calls. There is no queue or limit; every call is answered on the first ring.

What languages does Zanus AI support for on-call calls? The AI detects and responds in 40 languages, including Spanish, Mandarin, French and German. Your on-call technician speaks whatever language you have listed in their directory. If the caller speaks a language your tech does not, Zanus AI can announce the call anyway, and the caller and tech work it out, or you can route to a bilingual team member first.

If I do not like the trial, can I cancel and keep my phone number? Yes. You keep your number from day one. The only change is forwarding. When your trial ends or you cancel, you remove the forwarding and your number routes back to your old system. There is no porting or number loss.

Official sources and related guides

This independent Sysevo guide uses the following official documentation. The workflow and trial scenarios are evaluation proposals, not reported customer results or a hands-on benchmark.

• Zanus AI official product documentation
• Zanus AI official supporting guide
• Macro vs GoHighLevel for Service Teams: Choose the Working System
• Zanus AI HVAC Pilot: Score Repair and Estimate Call Outcomes
• Zanus AI for home services

Compare the same requirements with the relevant Sysevo product, Voice Agents and CRM. Review Sysevo plans and pricing or book a demo to discuss the proposed setup. Sysevo sells its own business software and is not affiliated with the competitors discussed here.