The future of ATO data and AI. Securely connecting with ChatGPT, Copilot and Claude.
The integration of AI tools with ATO data is revolutionising accounting practices, enabling rapid health checks and client insights. With secure connections like the MCP connector, practices can now streamline operations while ensuring compliance and data security.
Part of our AI in accounting coverage. See the full AI for accountants guide →
Connecting AI tools to ATO data has long felt off limits. With the right structure, authorisation and governance, it now takes minutes to build a practice health check, find clients at risk of ATO action and prepare a client meeting pack.
A year ago, I’d have told you AI was something to start investigating. That changed quickly in 2026 as new models landed. AI is now part of everyday life.
The question I hear most from accountants is whether it’s safe to use with client data from the ATO. Here’s how we’re answering it.
The ATO portal was the problem we set out to solve
Business Automation Works started with a conversation with a senior partner at a large practice about time, workload and repetitive tasks. That became ATOmate, which now helps more than 2,000 practices process ATO documents and client communications.
Practices then kept raising a different pain: logging into the ATO portal. Our research found the average accounting professional logs in five and a half times a day, with multi-factor authentication every time and a 20-minute idle timeout. Once inside, the data is slow to pull together, with no single view of a family group and no way to filter clients by partner.
So we built Omble, launched in 2026 and now used by around 150 practices. It brings ATO data such as lodgments, transaction history, balances and payment plans into one set of dashboards. It combines that with practice management data such as family groups, managing partners and relationship managers.
Getting ATO data into AI used to mean hours of downloads
Until recently, the honest answer to “can I safely use AI on my clients’ ATO data?” was no. The workaround was to log into the portal, download everything, build reports and drop them into an AI tool. That’s tedious, and you need to know what you’re looking for before you start.
With Omble already holding the data, we asked whether we could connect AI tools to it securely. We built an MCP connector that sits between Omble and your AI tool of choice. In simple terms, it controls the connection and who can access the data.
The data is the same live data you’d see in the portal, held securely in Australia. Because it’s structured, AI tools give far more consistent answers than they would from messy exports.
The practice controls who connects
Your practice authorises which users can connect, and each user must log in and pass multi-factor authentication. The connection only works when the user is authorised and the connector is approved in the practice’s AI account. The connector also keeps the AI tool working within set rails, searching the right data in a controlled environment.
Omble and ATOmate are both approved solutions with the ATO, and we’re audited every year against ATO, TPB and privacy requirements. We’re also ISO 27001 certified.
Set your governance before you connect
We’ve built the infrastructure. Governance is up to your practice.
My advice is to choose your AI tool first, set the governance, roll it out practice-wide and use a paid version. Paid versions keep your data within your subscription and out of model training.
Many practices go a step further. They roll out one tool and block the free versions, which removes the risk of a well-meaning team member putting client data in the wrong place. Only the tool your practice authorises can reach the connector.
Work that took days now takes minutes
I tested this on a masked copy of real practice data covering 1,550 clients.
1. Practice health check
At a practice level, one prompt returns total clients, total ATO debt, overdue lodgments and payment plans by managing partner and relationship manager, with director penalty accounts flagged. Built manually, that report takes hours, if not days.
2. Individual lodgments due before 15 May
A view of individual lodgments due before 15 May comes back as a visual chart, and surfaces returns with no partner or relationship manager assigned.
3. Clients at risk of ATO debt action
At a client level, the same data can flag risk early. A search for accounts over $100,000 with rising debt, general interest charges or penalties, plus overdue lodgments, returned 19 entities. Each came with its payment plan status, in a report ready for partners. That combination matters, because unpaid debt with missing activity statements is a key trigger for a director penalty notice. The AI also handled an imprecise prompt, reading “GDI” as GIC.
4. Family group meeting pack
The same applies to client meeting prep. A single prompt built a pack for a family group covering every entity, balance and lodgment, as a PowerPoint. The formatting needed a little work, though the content was ready. It also found a credit on one entity’s income tax account, a refund the firm could chase.
Would your team have gone to that level of detail before? In most cases, I hear the answer is no.
AI gives us time back with clients
People often ask whether AI will reduce headcount. I don’t believe it will. AI needs people to cross-check the data and apply judgement.
I see this as a back-to-the-future moment. This profession was built on client relationships and being a trusted advisor. Compliance pressure pushed much of that work behind closed doors. Tools like this give us the time to get back in front of clients.
We’re rolling the MCP server out to existing Omble users first, with training to help them get started.
Hours of logging in and collating become minutes. What you do with those minutes is the opportunity.
The Future of ATO Data + AI: Securely Connecting with ChatGPT, Copilot and Claude
Stan Corner, Founder and CEO of Business Automation Works, connects an AI tool to ATO and practice management data and runs four prompts, from a practice health check to a full client meeting pack.
Part of The AI in Practice Summit, 11 and 12 November 2026. All session recordings now available.
Get the editable AI policy template Download the AI policy template
The Word document (.docx), ready to fill in. Your profile unlocks every template on The Firm.
Or build it with AI
Open Claude or ChatGPT with the prompt filled in. It writes the finished document for your firm straight away, then asks a few questions to tailor it. No client data needed.
Works without client data. Don't paste client names, tax file numbers or ledgers into an AI tool your firm hasn't approved for them. AI data safety playbook.