The information is definitely in there. That is what makes it maddening. Every job you have ever run, every quote that went cold, every invoice still sitting unpaid: all of it is in Simpro, and most of it is three screens deep in a shape you cannot use. So someone exports to a spreadsheet. Or someone asks for a report, waits a week for it, and gets a number that was out of date the day it arrived.
We run a Simpro AI assistant for trade and field service contractors that changes the shape of that problem. You ask Simpro a question in plain English, from Microsoft Teams or from Claude, and you get the answer from the live system, scoped to what you are allowed to see. Then you tell it what to do next. This post is about what that looks like in practice, what it can and cannot do, and why the way it is built matters more than the fact that it exists.
Simpro is good software. It runs quotes, jobs, scheduling, materials, invoicing, assets and recurring maintenance contracts for thousands of Australian contractors, and it is heavily used across WA. We are not here to knock it.
The difficulty is structural. A job management system is built to record work as it happens, one record at a time. It is not built to answer the questions a business owner asks on a Monday morning, which are almost always across records: which of my open jobs are losing money, which quotes has Dave let go quiet, who owes me more than they should. Simpro’s own reporting covers some of this. The moment your question is slightly different from the report that exists, you are back to exporting.
The usual fix is a business intelligence project. Copy Simpro into a data warehouse, put Power BI on top, build dashboards. That works, and plenty of contractors do it. It also costs money, takes months, needs someone to maintain it, and answers the questions you thought of when you built the dashboard. It does not answer the one you thought of this morning.
Strip away the marketing and the idea is simple. We connect Claude, Anthropic’s AI model, to your Simpro instance using an open standard called the Model Context Protocol. Claude can then read from Simpro, and with your approval write to it, in response to a question typed in ordinary English.
Three design decisions make our version different from a chat window bolted onto a spreadsheet export, and they matter enough that we will keep coming back to them.
It reads the live system. Not last night’s copy. When you ask where job 4182 is at, the answer reflects the note a technician added twenty minutes ago.
It acts as you. The connection carries your Simpro permissions. A technician sees their jobs. A branch manager sees their branch. The owner sees everything. Nobody gets a pooled copy of the whole database because they happened to have the chat window open.
It runs where you already are. Microsoft Teams, or the Claude app. There is no new dashboard to open and forget about, and no new login for the office to learn.
The managed AI service we wrap around it handles the connector, the access model, the logging of every question asked, and the ongoing tuning. That last part turns out to be most of the work, and we will explain why further down.
Here is the part most people actually want to read. These are real questions the assistant answers today, grouped roughly the way a contractor’s week is grouped.
“Where is job 4182 at?”
You get the customer and site, the current stage and status, the due date, value ex and inc GST, cost centre totals against what has been claimed so far, what is scheduled next, open tasks, recent notes, and a flag on anything that looks at risk. With a link straight to the job in Simpro so you can check it yourself.
“Is job 4182 actually making money?”
Estimated cost versus actual, committed costs still to land, revenue, gross margin, how much has been claimed versus not, and which cost centres have gone over budget. This is the Simpro job profitability question every contractor wants answered per job and usually only sees in aggregate, months later.
“Show me every job over $50k still open, worst margin first.”
“What is in the quote pipeline?”
Open quotes totalled by stage, by salesperson and by customer, with ageing buckets and a list of the oldest ones still sitting unanswered.
“Which quotes has Dave got out that have not moved in three weeks?”
“What is our quote-to-job conversion this quarter compared to last?”
“Who owes us money?”
Unpaid invoices grouped by customer and aged 0–30, 31–60, 61–90 and 90+ days, with the top debtors, the oldest invoices, and a link to each one. The PDF is one click from there. Finding overdue Simpro invoices stops being a job and becomes a sentence.
“How much have we invoiced each month for the last year?”
“What did materials cost us this quarter versus last?”
“What has Dave got on Thursday?”
A chronological day plan: time blocks, site, customer, job name and any notes attached.
“Which techs are booked solid next week and which have gaps?”
“Show me overdue maintenance jobs grouped by technician.”
“Give me everything on Westland Facilities before I call them.”
Who they are, their sites and key contacts, open jobs and quotes with values, unpaid invoices and total outstanding, and anything overdue. One screen, with links to every record named. We have watched account managers who used to spend ten minutes clicking around before a difficult call do this in the car park.
“Find that job with the switchboard upgrade on Hay Street.”
It searches customers, contacts, sites, the catalogue, vendor orders, tasks and assets from a fragment. A part number, half an address, a reference someone half-remembers.
“What has changed since Friday?”
Everything modified across jobs, quotes, invoices and customers since a date, grouped and counted, with the stage and status changes called out. This is a properly good Monday morning question, and no report we have ever seen a contractor commission answers it.
A reporting tool tells you the quote has gone quiet. Then you open your email and write the follow-up. That gap, between knowing and doing, is where most of the time actually goes.
So the assistant does the next thing too.
“Draft the follow-up emails for every quote that has gone quiet for three weeks.”
“Add a note to job 4182 that the client has approved the variation.”
“Raise the invoice for the completed stage on job 4182.”
“Schedule Dave for the Hay Street job Thursday morning.”
The pattern throughout is the same: it prepares the work, and a person approves it. Drafts come back for review before anything is sent. Records are shown before anything is written. We are not interested in building an assistant that raises an invoice on its own initiative, and we do not think you should be interested in buying one. The value is in having the ten minutes of preparation done, not in removing the person who signs it off.
Everything above is useful. None of it is the reason we built this.
The questions a contractor most wants answered are the ones that need two systems at once, and today those questions are answered by a person and a spreadsheet or not at all. Simpro knows what you quoted and invoiced. Xero knows what actually landed in the bank. Nobody has both open at the same time, so nobody knows the gap.
“Reconcile this month’s Simpro invoices against what has actually landed in Xero and show me the differences.”
“Which jobs have we invoiced but never been paid for?”
“Job 4182’s material costs. Do they match the supplier bills we have been charged for?”
“What did we quote this month, what did we invoice, and what did we bank?”
That last one is the question every contractor wants and nobody can produce, because the three numbers live in three places. Once both systems are connected, it is one sentence.
The same logic extends to the rest of the tools your office already runs. With email and Teams: “The client has emailed about the Hay Street job. Pull it up, tell me where it is at, and draft the reply.” Or a standing instruction: “Every afternoon, post tomorrow’s schedule to the ops channel with any job that is missing a tech.” With SharePoint: “Build the compliance pack for this site, test certificates for every asset plus the job history.” With the phone system: “Who called about job 4182 last week and what happened?”
Every one of those pairings runs today: Simpro alongside Xero, Microsoft 365 (email, Teams and SharePoint) and the phone system. Once two systems are connected the same way, questions that span them work without a project. If your office runs something we have not named, a Deputy roster or a supplier portal say, that is a build rather than a blocker, and usually a short one.
There are good Simpro reporting products out there. Most of them work the same way: extract Simpro into a data warehouse, connect Power BI on top, and increasingly offer an AI connection to that warehouse as well. If what you want is a set of dashboards the whole business watches all year, that is a reasonable way to get them, and we would not talk you out of it.
The difference is not that one is good and one is bad. It is that they are different products solving different problems, and it is worth being clear about which problem you have.
| Warehouse and dashboard tools | Our Simpro AI assistant | |
|---|---|---|
| How current is the answer | As at the last sync | Right now, from the live system |
| Who sees what | Whoever can open the warehouse sees the copy | Each person sees their own jobs, customers and companies, per their Simpro permissions |
| What it does | Reports and dashboards | Answers, then does the next thing: drafts the email, raises the invoice, adds the note |
| Where you use it | Power BI, or the vendor’s app | Microsoft Teams and Claude |
| Multiple Simpro companies | One pooled copy | Scoped per person, per company |
| Who sets it up and tunes it | You, self-serve | Us, hands-on, as part of a managed service |
If you run a multi-company Simpro setup, the “who sees what” row is the one to think hardest about. A pooled warehouse means the person running reports for company A can, in principle, see company B. An assistant that acts as you cannot. For Simpro multi-company reporting that distinction is the whole story.
Here is the thing we have learned doing this work that we wish someone had told us earlier. The connection is the easy part. Anyone can wire an AI model to an API now; the standards are open and the tooling is good.
The hard part is that every business system has places where the data quietly lies. A status that half the office uses one way and half another. A cost centre that was renamed in 2023 and still has old jobs against the previous code. Archived customers that still have open invoices. A report view that excludes cancelled jobs, which is right for one question and wrong for another.
Point a generic AI at a system like that, ask a simple question, and you will get a confident, specific, wrong number. It will look completely reasonable. Nobody will catch it, because the whole point was not having to check.
Making a Simpro AI assistant trustworthy means measuring the data against itself, writing down where it misleads, and building those traps into how questions get answered. When ours reports job profitability, it says which costs are committed but not yet landed, because a margin that ignores those is fiction. When it counts overdue invoices, it tells you whether it has included disputed ones. That is the work a client is paying an integration specialist for, and it is precisely what a self-serve connector cannot give you, because it does not know your business.
Contractors with somewhere between 10 and 200 staff, running Simpro, in electrical, plumbing, HVAC, fire protection, security or facilities maintenance. Businesses that have years of data in the system and a standing frustration about getting it out. Usually there is one person in the office who has become the unofficial Simpro report writer and would like their Fridays back.
We are a Perth managed IT provider and we deliver this hands-on, which means WA contractors are who we work with most. The connector itself does not care where you are, so if you are running Simpro in Adelaide or Brisbane and the questions above are your questions, talk to us and we will tell you honestly whether we can deliver it well for you.
Two things to be aware of before you enquire. First, this sits inside a broader AI services engagement rather than being a box you buy, because the tuning and the access model are where the value is. Second, we are careful about what “connected to AI” means for your data: every question is logged, permissions follow the person, and nothing writes to Simpro without approval. If a provider cannot explain those three things to you in plain English, that is a gap.
Write down the five questions you cannot currently get out of Simpro. Not the ones a dashboard already answers. The ones where you sigh and open Excel. Job margin per job, quote conversion by salesperson, invoiced versus banked, whatever they are for you. That list is the brief, and it tells you within a minute whether an assistant like this is worth your time.
Check who can see what today. Open Simpro and look at how permissions are set for your technicians, your office staff and your managers. If everyone is an administrator because that was easier, fix that first. Any AI connection inherits those permissions, so it will only be as careful as your Simpro setup is.
Watch it answer one of your questions. We have a working demonstration against a realistic contractor’s Simpro dataset. Bring your list, pick one, and watch it come back with the answer and the links to check it. That takes about twenty minutes and commits you to nothing. Contact us on 1300 EPIC IT or through the contact page to book it.
Yes. A Simpro AI assistant connects an AI model such as Claude to your Simpro instance so you can ask questions in plain English and get answers from your live data. Ours also acts on the answer with your approval, for example drafting quote follow-ups or raising an invoice, and runs inside Microsoft Teams or the Claude app.
Yes. Power BI is the right tool for standing dashboards the whole business watches, but most of the questions contractors actually ask are one-off. A Simpro AI assistant answers those directly from the live system without building a report first, and links each answer back to the records in Simpro so you can verify it.
Ask for them. “Who owes us money?” returns unpaid invoices grouped by customer and aged into 0–30, 31–60, 61–90 and 90+ day buckets, with the top debtors and a link to each invoice PDF. You can then ask it to draft the reminder emails for anyone over 30 days and review them before they go.
It depends entirely on how the connection is built. Ours reads the live system as the person asking, so each user sees only what their Simpro permissions allow. Every question is logged, and nothing is written back to Simpro without a person approving it. Ask any provider to explain those three points before you agree to anything.
Yes, and this is where the live-connection approach matters most. Because the assistant acts as the person asking, someone with access to one company cannot see another’s jobs or invoices. Warehouse-based Simpro reporting tools typically hold a single pooled copy of all companies, which is a different security posture.
Perth and WA are where we deliver most of our work, because we set this up hands-on rather than self-serve. The connector itself works anywhere in Australia. If you are outside WA, contact us and we will tell you plainly whether we can deliver it well for your business.
