AI Receptionist with ServiceM8 and Jobber: How It Works
"Does it work with ServiceM8?" or "Does it plug into Jobber?" is usually the first question a trade business asks about an AI receptionist. It is a good question, and it is also easy to answer too quickly.
A receptionist that answers calls is useful. One that also puts a clean job request into the software your team already uses saves real admin time. But "integrates with" covers a lot of ground, from a full two-way sync to an email that someone still has to retype. This article explains how these connections work in general, what the vendors actually document, and what to ask before you trust a claim.
We cover two products because they are common with trades: ServiceM8, popular with Australian tradies and contractors, and Jobber, common with US home-service teams. We only describe integrations that a vendor documents in public, and we say where we could not verify something.
The basic idea
When an AI receptionist takes a call, it collects a handful of facts: who is calling, a phone number, an address or suburb, what is wrong, and when they want someone out. The integration's job is to turn those facts into a record in your job software.
There are three common ways it happens.
1. Built into the software. The software vendor makes the receptionist itself. Jobber's Receptionist is an example. The data stays in one system, but you use the receptionist the vendor built, not one tailored to you.
2. A direct connection through the software's API. A third-party receptionist calls the job software's API to create a client or a job. This is more flexible, and the quality depends on how much of the API the receptionist uses.
3. A bridge through an automation tool. The receptionist sends the details to a workflow tool, which then creates records in the job software. This is the most adaptable and the easiest to adjust, but there is an extra moving part to keep running.
All three can work. What matters is exactly what gets created, and what still needs a person.
What Jobber documents
Jobber's help page describes a Receptionist that answers calls and texts around the clock, books visits, takes requests and escalates to you when needed. When a job cannot be booked outright, the page says it gathers the details and "creates a work request for you to review". It says it can book a visit during the call using your existing online booking settings, and that new leads and appointments flow into your schedule and client list. Conversation logs include a short summary, a full transcript and a recording.
Three things the page does not say, as of 10 November 2026:
- It lists no price and does not say which plans include the Receptionist.
- It does not describe disclosure or consent wording for callers, which is your job. See call recording consent for AI receptionists.
- It does not explain how a third-party receptionist would write into Jobber. We tried to read Jobber's developer documentation and the page refused an automated request, so we are not going to describe Jobber's API for outside tools.
The page also describes a companion feature that would draft quotes during calls and marks it as coming soon. We do not count features that are not yet available.
If you are a Jobber customer, the simplest route is often to try Jobber's own Receptionist first and see whether it covers your calls. A custom build makes sense when you need rules, wording or routing that the built-in one does not offer.
What ServiceM8 documents
ServiceM8's developer portal gives a short overview: it describes building add-ons as creating "serverless add-ons that extend the ServiceM8 app and online platform", and it has an "Integrate" section about connecting apps to cut duplicated manual work. The page itself has no technical detail, and we could not open the reference pages behind it. So we cannot confirm from ServiceM8's own pages what the API allows.
What we can point to is a third party's documentation of a connection. ElevenLabs documents a ServiceM8 integration for its Receptionist product. According to that page:
- You create an API key in ServiceM8 under Settings, then API Keys, and the key is sent in an
X-Api-Keyheader. - One tool creates a job when a caller needs work done, with a status of Quote or Work Order, and it needs an address and a description.
- Another tool adds the caller as a new client with a name and optional address.
- Both return only "OK", so the new record's ID is not available to the assistant.
- Phone numbers and emails live on a separate contacts record that the client tool cannot create, so the phone number goes into the job description instead.
- The assistant cannot look up existing clients.
- For linked clients and jobs, the page suggests using Zapier's Create Client and Create Job actions instead.
That is a useful, honest picture. It shows what a working connection looks like and where it stops. It is also one vendor's setup, not a statement about everything that is possible with ServiceM8. A different build could do more, and we would need to confirm the details in ServiceM8's reference documentation before promising it.
The gaps that matter in real life
From those two examples, here are the practical gaps to ask about.
Existing customers. If a repeat customer calls, does the system recognise them, or create a duplicate? Duplicates are annoying at best and confusing for a team at worst. The ServiceM8 example above says the assistant cannot look up existing clients, so someone has to merge them later.
Linking the job to the client. If the client and job are created separately and the IDs are not passed back, a person has to connect them. That is a small task per call, but it is the kind of small task that grows.
Where the phone number goes. Putting it in the job description works, but it is not the same as having it on the contact record where your texting or calling tools look for it.
Status and review. Jobber's flow creates a work request for you to review, which is a sensible default. For ServiceM8, a job created as a Quote is something your team still has to price and follow up. Decide up front who looks at new items, how often, and what counts as urgent.
Availability and double-booking. An assistant that books a visit needs to know your real calendar. Jobber says it uses your existing online booking settings. For any other setup, ask how it knows who is free, and what happens if two callers want the same slot.
Failures. What happens when the connection breaks? An API key expires, a field changes, a service has an outage. Ask whether the call details are kept somewhere else as a backup, and whether you get an alert.
Questions to ask any vendor
Put these to a vendor, including us:
- Which of your integrations do you document publicly? Can I see the page?
- Exactly which records are created: client, contact, job, request, appointment?
- Which fields are filled, and which are left for my team?
- Does it recognise a customer who has called before?
- Does it write back anything I can see, such as an ID or a link to the call recording?
- What happens when the connection fails, and how will I know?
- Who maintains it if the software changes its API?
- Can I switch the booking step off and have it only create a request?
A vendor with clear answers is worth more than one with a long feature list.
How we approach it
We build AI receptionist and booking flows around the software a trade business already runs. During the audit we confirm which systems and channels your setup supports, because the options differ between Australia and the US and between plans. We would rather tell you that a certain step stays manual than promise a full sync we cannot show you documentation for.
A sensible first version is modest: the AI collects the details, creates a request or job in your software with the contact details in it, and tells your team. You review it. Once that is working and trusted, you widen what it can do, such as checking availability or booking slots. This is also easier to explain to customers and easier to audit.
What to do next
If you use ServiceM8 or Jobber today:
- Write down the last ten calls you missed or handled badly, and what you would have wanted to happen to each.
- Check whether your software vendor offers its own receptionist, and what it covers.
- List the exact fields you need in a new job: name, number, address, problem, preferred time, urgency.
- Decide who reviews new requests and by when.
If you want help working out which route fits, that is what the AI audit covers. And if you are still weighing the bigger choice, read AI receptionist versus answering service first.
Sources
- ServiceM8, ServiceM8 developer portal, undated. Accessed 2026-11-10. High-level overview only. It describes add-ons as serverless extensions of the ServiceM8 app and platform and points to separate documentation.
- ElevenLabs documentation, ServiceM8 integration (Receptionist AI), undated documentation page. Accessed 2026-11-10. Third-party vendor documentation of one ServiceM8 connection. Not ServiceM8's own documentation.
- Jobber, Jobber Receptionist, undated help page. Accessed 2026-11-10. Jobber's description of its own built-in AI receptionist.