We take one manual process, such as supplier invoices keyed into Xero or a shared inbox sorted by hand, and automate it inside the systems you already run. Each phase is quoted at a fixed price. Nothing is switched over until the automation has run alongside your team and matched their work.
Process automation fits when the same work arrives every day in roughly the same shape and someone retypes it from one place into another. If any of these sound familiar, it is worth a conversation.
Invoices, delivery dockets, forms and statements arrive as PDFs and someone types the numbers into Xero, MYOB or an ERP.
The CRM, the accounts package and the job system each hold part of the picture, and spreadsheets fill the gaps between them.
A shared mailbox gets hundreds of messages and a person reads each one to decide who should deal with it.
A wrong amount or a duplicated bill gets caught at month end, or by a supplier, and someone has to trace it back.
The same report is rebuilt every week or month by exporting, pasting and reformatting the same data.
Nobody can say how many items are waiting, how long they have waited or where they get stuck.
These are the processes buyers most often bring to us. Each card links to a longer guide on how that kind of automation works.
Read invoices in any layout, extract supplier, ABN, GST and line items, check for duplicates and create draft bills in Xero or MYOB for approval.
Read guidePull data from contracts, forms, dockets and PDFs, check it against your rules and push it to the system that needs it.
Read guideRead incoming mail, classify it by intent, draft a reply where that is safe and route the rest to the right person.
Read guideMatch bank lines to bills and invoices, and send the items that do not match to a bookkeeper with the likely candidates attached.
Read guideBuild the weekly or monthly report from live data, with the same structure every time and a log of where each figure came from.
Read guideCollect documents, check ABNs and bank details, create the records and chase whatever is missing.
Read guideThis is the typical shape of a first phase that automates one process, and your own timeline may differ. A process that touches several systems, or data that needs cleaning first, takes longer, and the timeline for your work is written into the proposal after discovery.
We sit with the people who do the work and time it on real items. Estimates from memory are usually wrong, so we measure. By the end of the week we have numbers for:
You receive a process map, the baseline figures, a list of the exception types we found and a written recommendation. If the numbers support a build, the recommendation includes a fixed price for it. If they do not, it says so and suggests what would help instead. This is the same discipline as our four-week proof of concept framework and our automation business case template.
We connect to your systems through their official APIs, build the extraction and the business rules, and set up three things every automation needs: an exception queue where a person reviews anything the automation is unsure about, a log of every action it takes, and alerts when something fails. Before the shadow run starts we test against a set of your past items where the correct answer is already known. You see progress each week in a short demo using your own data. Where the work needs two systems connected first, that piece follows our system integration approach.
In shadow mode the automation runs alongside the people doing the work. Your team keeps processing items the normal way. The automation processes the same items at the same time, but its output goes to a holding area and never reaches your live systems.
Each day we compare the two results item by item. Where they differ, we find out why. Sometimes the automation misread a document and the rule gets fixed. Sometimes the person made the mistake, which is useful to know too. The switch-over happens only when the match rate reaches the threshold you agreed to in week 1, and you see the comparison report before you decide.
We switch over in stages, usually starting with the item types that matched most reliably in the shadow run. Anything else keeps going to a person. The manual process stays documented so your team can resume it the same day if the automation is ever paused.
At the end of week 4 you can see how many items the automation handled, how many went to a person and the reason for each, set against the baseline from week 1. Your team has been trained on the exception queue, you hold the runbook, and you make a go or no-go call on whether to extend to a second process. If you would rather we keep running and monitoring it, that is our managed AI service.
See what this would take in your business
A free 30-minute review on Microsoft Teams. Bring the process you want fixed and leave with a clear first step.
Picture a business that receives its supplier invoices as PDF attachments to an accounts inbox. An accounts payable officer opens each one, checks the supplier and ABN, types the amounts, GST and account codes into Xero, matches it to a purchase order where there is one, and saves the bill for approval.
Whether 47 hours a month justifies a build depends on the fixed price in your proposal, the running cost and what those hours are worth to you. The straight-through rate is the assumption that moves the answer most, which is why it is measured in the shadow run before anyone relies on it. You can try your own figures in the ROI calculator or the payback period guide, and our honest guide to invoice automation covers where these projects go wrong.
| Metric | Today | After go-live |
|---|---|---|
| Opening each PDF | Every invoice | Exceptions only |
| Typing amounts and GST | Every invoice | None, values are extracted |
| Duplicate checks | By memory or search | Automatic on every invoice |
| Approval in Xero | Manager approves | Manager still approves |
| Audit trail | Email folder | Log from PDF to bill |
For the technical detail of how extraction and Xero posting fit together, see how AI invoice processing works.
Everything below is yours at the end of the phase, including the code and configuration.
You pay for phase one, you own what it produces, and you can stop there.
Phase one is quoted at a fixed price after a discovery call. Process automation starts from $5,000, and the full list of starting prices is on our pricing page. Every phase ends with a go or no-go decision. If you stop, you keep the process map, the baseline, the documentation and whatever has been built.
If discovery shows the numbers do not stack up, we tell you before you spend more.
Some processes should not be automated yet, and some problems need a different kind of work. We would rather say so at the start.
Rules that keep moving cost more to maintain than they save. Let the process settle first.
A task done twenty times a month rarely pays for a build. A template or checklist is usually enough.
Automating a bad process makes the mistakes faster. Sometimes the cheaper engagement is fixing the steps, which is what our IT consulting work covers.
If no one is willing to review a draft before it goes out, the automation has nowhere to send its uncertain cases. Keep that work with people.
That is an integration job. Once the data flows, the manual step often disappears without any AI at all.
When the gap is a tool you do not have yet, custom software is the better route.
See what this would take in your business
A free 30-minute review on Microsoft Teams. Bring the process you want fixed and leave with a clear first step.
Based in Brisbane, working with businesses across Australia, with Australian hosting available.
Scope and price are agreed before work starts. No hourly billing.
Our principal consultant's prior project experience includes BHP, Rio Tinto and Senex Energy, gained with previous consulting employers. These companies are not Solve8 clients.
Our carbon reporting work for a Tier 1 infrastructure contractor used the same discover, build and hand over approach.
We use the tools that fit your systems and have no platform to sell you.
Documentation and training are part of every phase, so your team can run what we build.
Read the Tier 1 carbon reporting case study or see how it applies to construction and infrastructure.
Our principal consultant brings 18 years of enterprise delivery experience from projects for BHP, Rio Tinto and Senex Energy, delivered with previous consulting employers. These companies are not Solve8 clients. You work directly with senior people, not a sales team handing you off to juniors.
Book a free 30-minute call. Tell us which process takes the most hands-on time, and we will tell you whether automation is likely to pay and what the first phase would involve.