Connecting a CRM to ServiceM8: Custom Build vs Odoo, Salesforce, HubSpot, Zoho or Monday.com

ServiceM8 is excellent at what it was built for: getting a job from enquiry to quote to schedule to invoice without dropping information along the way. What it is not is a full customer relationship management (CRM) system. It doesn't do sales pipelines the way a dedicated CRM does, it doesn't do marketing automation, and it doesn't give your sales team a single view of every touchpoint a customer has ever had with your business.

So a very common question from growing Sydney service businesses is: how do I connect a proper CRM to ServiceM8? You want ServiceM8 to keep running the field-service side — jobs, technicians, invoices — while a CRM handles leads, deals, quoting workflows, and marketing. Or you want to build a CRM that fits your business exactly, rather than bending your process to fit someone else's software.

There are two honest answers, and both take real engineering. This article walks through both paths, what each of the major CRM platforms actually exposes through its API, why "merging the data" is the hard part, and how we approach these builds at Proanalytica — including using an AI agent like Hermes to hold the whole thing together.

First, the uncomfortable truth: there is no plug-and-play connector

If you go looking for a native, one-click ServiceM8-to-Salesforce or ServiceM8-to-HubSpot button, you won't find one. ServiceM8's built-in integrations are focused on accounting (Xero, MYOB, QuickBooks), payments, and a handful of add-ons. It does not ship with a maintained connector to the big CRM platforms.

What ServiceM8 does have is a comprehensive REST API. Nearly everything you can do in the ServiceM8 interface — jobs, companies (clients), job contacts, quotes, invoices, materials, staff, schedules — has an API equivalent. It also supports webhooks, so ServiceM8 can notify an external system the moment a job status changes, a quote is accepted, or a client record is updated. Authentication is via API key or OAuth for registered add-ons.

That means every CRM connection is a build, not a setting. You are writing (or having someone write) a custom connector that talks to two APIs, decides what "the same customer" means in both systems, and keeps them in sync without creating duplicates. The good news is that this is a well-trodden path — the bad news is that every "just connect them" tool you'll see advertised is really doing this work under the hood, usually with less control than you'd get from a purpose-built connector.

Path one: build a custom CRM on top of ServiceM8

For a lot of trades and service businesses, the honest best answer is not to buy a CRM at all. It's to build a lightweight, custom CRM layer that uses ServiceM8 as the operational source of truth and adds exactly the sales and relationship features you're missing — no more, no less.

This works because ServiceM8's API is genuinely capable:

• The data model is consistent and well-structured. Companies, contacts, jobs, quotes and invoices come back as predictable JSON, which makes building a front-end against it far easier than with older platforms.
• Webhooks give you near-real-time updates. When a quote is accepted or a job is completed in ServiceM8, your custom CRM can react within seconds instead of polling on a timer.
• Your team doesn't change how they work. Technicians keep using the ServiceM8 app in the field. The custom CRM is a layer on top for the office and sales side.

A custom CRM build typically adds a proper sales pipeline (leads and deals with stages ServiceM8 doesn't model), follow-up reminders and task tracking, a unified customer timeline that stitches together every job, quote and invoice for a client, and reporting that answers questions ServiceM8's stock reports can't — like win rates by suburb or average time from enquiry to accepted quote.

When this is the right call: when your sales process is specific to your business, when a full CRM licence per seat feels like paying for features you'll never use, and when you want your data to stay in systems you control. We've built exactly these layers for Sydney service businesses across electrical, plumbing, HVAC, property services and commercial cleaning.

Path two: connect an existing CRM

Sometimes you already have a CRM, or you have good reasons to use an established platform — your sales team knows it, it has marketing automation you rely on, or your accountant/ops stack is already built around it. In that case the job is to connect it to ServiceM8 through a custom connector.

Here's what each of the common platforms actually gives you to work with. This matters, because the shape of each API dictates how the connector is built.

Salesforce

Salesforce exposes a mature REST API (alongside SOAP, Bulk 2.0 for large data volumes, and Streaming APIs for change events). Authentication is OAuth 2.0 through a connected or external client app. You query records with SOQL and read/write the standard objects — Accounts, Contacts, Leads, Opportunities — plus any custom objects and fields you've defined. It's the most enterprise-grade option, and also the one where a connector has the most surface area to get right: field-level security, validation rules, and governor limits all shape what the connector can do. A ServiceM8 "company" typically maps to a Salesforce Account, a ServiceM8 job contact to a Contact, and a ServiceM8 job or accepted quote to an Opportunity.

HubSpot

HubSpot's CRM API is clean and modern. Everything is an object with a numeric type ID — Contacts (0-1), Companies (0-2), Deals (0-3), Tickets (0-5) and so on — and you create, read, update and search them through consistent /crm/objects/... endpoints. It has a strong Associations API for linking records (a contact to a company to a deal), a Search API for filtering, and built-in deduplication using unique identifiers like a contact's email or a company's domain. That dedup behaviour is genuinely helpful when syncing from ServiceM8, because it gives the connector a natural key to match on. HubSpot is often the smoothest CRM to pair with ServiceM8 for a small-to-mid business that wants marketing automation on top.

Zoho CRM

Zoho CRM offers a REST API with OAuth 2.0 authentication and modules that map neatly to the sales world — Leads, Contacts, Accounts, Deals. It's popular with cost-conscious Australian businesses because the licensing is affordable and the API is capable enough for a solid two-way sync. The connector work is similar in shape to HubSpot: match on email or a custom ServiceM8-ID field, then create or update the corresponding Zoho module records.

Odoo

Odoo is different, and worth flagging. Its external API is XML-RPC (not a modern REST/JSON API), accessed through xmlrpc/2/common for authentication and xmlrpc/2/object for everything else, calling generic ORM methods like search, read, search_read, create and write against models such as res.partner. Two practical gotchas: external API access is only available on Odoo's custom (self-hosted or paid) plans, not the free or standard tiers; and because Odoo is a full ERP, "CRM" data lives alongside sales orders, inventory and accounting, so the connector has to be deliberate about which models it touches. When it fits, though, Odoo lets you consolidate CRM and back-office in one place, with ServiceM8 feeding the field-service reality into it.

Monday.com

Monday.com (including its Sales CRM product) runs on a GraphQL API. Instead of fixed endpoints you query for exactly the boards, items, columns and users you need, and mutate them the same way. Authentication is via API token or OAuth. Because Monday is board-and-item based rather than a traditional record store, mapping is a design decision: a ServiceM8 client might become an item on a "Customers" board, a job an item on a "Jobs" board, with column values carrying status, value and dates. It's flexible and visual, which suits teams that want their CRM to look like a workflow rather than a database.

The hard part isn't the connection — it's merging the data

Connecting to two APIs is the easy 20%. The other 80% is deciding what happens when the same customer exists in both systems with slightly different details, and keeping them agreeing over time. This is where naive integrations fall apart and start spawning duplicates.

The problems you have to solve for any of these platforms are the same:

• Identity matching. ServiceM8 doesn't share an ID with your CRM. So how do you know that "J. Smith Plumbing" in ServiceM8 is the same record as "J Smith Plumbing Pty Ltd" in Salesforce? You need a matching strategy — usually email or phone as a natural key, backed by a stored cross-reference ID once a match is confirmed, so you never have to guess twice.
• Direction and conflict rules. If a phone number changes in both systems between syncs, which one wins? A robust connector defines a source of truth per field — for example, job and invoice data flows from ServiceM8, while sales stage and marketing consent flow from the CRM.
• Deduplication. Without a match key, every sync creates a new record. HubSpot's built-in dedup helps; Salesforce, Zoho, Odoo and Monday.com mostly leave it to you. This is the single most common reason DIY integrations get abandoned.
• Rate limits and pagination. ServiceM8, Salesforce and Monday.com all enforce API limits. A connector that tries to pull everything on every run will get throttled. Webhooks plus incremental, changed-record syncing is the pattern that scales.
• Field and unit mismatches. Statuses, currencies, GST treatment and picklist values rarely line up one-to-one between two systems and have to be mapped explicitly.

This is why the deliverable is almost always a small middleware layer — a lightweight Node.js service, or an automation platform like n8n — sitting between ServiceM8 and the CRM. It listens for ServiceM8 webhooks, applies the matching and mapping rules, calls the CRM's API, and logs everything so you can see exactly what synced and why. It's not glamorous, but it's the difference between a sync you can trust and a growing mess your bookkeeper has to untangle every month.

Then the fun part: dashboards and custom websites on top

Once the data is flowing reliably between ServiceM8 and a CRM, you have a unified dataset that neither system presents well on its own. That's where the visible value gets built.

Custom dashboards. ServiceM8's stock reports and a CRM's default dashboards each show half the picture. A custom dashboard, pulling from both, can show the whole funnel — leads in the CRM, conversion to accepted quotes, jobs completed in ServiceM8, revenue and outstanding invoices — in one branded view. We've built these for clients so an owner can open one screen and see win rates by suburb, technician utilisation, and cash owed, all live.

Custom client-facing websites. We've also built custom websites that sit on top of this combined data: branded client portals on the business's own domain (not a ServiceM8 subdomain), live project trackers for commercial and strata clients, and public job showcases that turn completed work into a sales asset. The website reads from ServiceM8 and the CRM through the same connector, so nothing is ever entered twice.

The common thread across every one of these builds: the operational systems stay as the source of truth, and the presentation layer — dashboard or website — is custom-built to fit the business. ServiceM8 keeps running the jobs, the CRM keeps running the pipeline, and the custom layer makes the combined data actually usable.

Tying it together with an AI agent like Hermes

The final piece is what turns a set of connected systems into something that feels intelligent. Once ServiceM8 and your CRM are talking through a reliable connector, you can put an AI agent — like Hermes or Openclaw — on top of the unified data to do the work that rules-based automation can't.

A few of the things an AI agent does well in this setup:

• Resolving the messy matches. Deterministic matching handles 90% of records cleanly. For the ambiguous 10% — "is Smith Property Group the same as Smith Properties?" — an AI agent, given both records as context, can make the call that a strict rule can't, and flag anything genuinely uncertain for a human.
• Enriching and summarising. The agent can read a customer's entire history across both systems — every job, quote, note and email — and produce a plain-English summary before a sales call, or draft the follow-up.
• Natural-language reporting. Instead of building a new dashboard tile every time someone has a question, you can ask the agent: "which commercial clients haven't been quoted this quarter?" and it queries the combined ServiceM8 and CRM data to answer.
• Triage and routing. New enquiries can be classified, matched against existing customers, and routed to the right person or pipeline stage automatically.

The important principle — the same one we apply to the connector itself — is that the AI agent is used where judgement is needed, not as the primary path for everything. The deterministic connector moves the clean data. The agent handles ambiguity, summarisation, and the natural-language layer on top. That hybrid is what makes it dependable rather than a black box.

What this looks like in practice

Every one of these builds is scoped to the business. A sole trader doing ten jobs a week probably doesn't need any of this — ServiceM8 on its own is plenty. But if you're running a growing service business in Greater Sydney with a real sales pipeline, commercial or strata clients who expect polished reporting, and a team that's tired of entering the same customer into two systems, connecting a CRM to ServiceM8 properly pays for itself.

At Proanalytica Technologies we do this end to end: choosing between a custom CRM and an off-the-shelf platform, building and maintaining the connector, handling the data merging and deduplication, building the dashboards and custom websites on top, and layering in an AI agent like Hermes to tie it together. We've done it for clients across electrical, plumbing, HVAC, property services and commercial cleaning, and the architecture adapts to each — the principle stays the same.

Get in touch to talk through what connecting a CRM to your ServiceM8 setup would look like for your business.