How we use AI at ITAAAS

Version 1.1, 11 October 2026. 

What we use AI for

We use AI tools as assistants for defined, bounded tasks:

• Summarising meeting recordings, transcripts, documents and research, so our consultants spend their time on analysis rather than note-taking.
• Automating repeatable administrative tasks, such as drafting the CRM update after a client meeting or preparing follow-up actions, which a consultant reviews before anything is saved or sent.
• Converting content between formats, for example turning a written report into a presentation, or a set of notes into a structured document. The same applies to our own marketing, podcast and newsletter material.
• Challenging our work. Once we have formed a position and tested it internally, we ask an AI tool to review it, argue against it, and find what we've missed.
• Helping write code, scripts and configuration in delivery and build work, where an engineer reviews, corrects and tests everything before it is deployed.

What we don't use AI for

We do not use AI to do our thinking for us. Specifically:

• AI does not originate our advice. Every ITAAAS recommendation starts with our own analysis and experience.
• AI does not replace peer review. Our work is challenged by senior colleagues before any tool sees it.
• Nothing goes to a client unreviewed. A named ITAAAS consultant or engineer is responsible for every deliverable, and that responsibility is never delegated to software.

The order we work in

For advisory work the sequence matters more than the tools, so here it is:

1. We form our view first. Before any AI input, we write down our own analysis, options and recommendation.
2. We review it internally. Senior colleagues challenge the reasoning, the assumptions and the trade-offs.
3. Then, and only then, we ask AI to challenge it too. The tool acts as a further reviewer, never the author.
4. A named consultant signs off. They own the final deliverable and answer for it.

Build and delivery work

Where we build or configure systems, the order is different and we'd rather say so than pretend otherwise. AI coding assistants often produce the first draft of code, scripts or configuration. A named engineer reviews, corrects and tests all of it before deployment and is accountable for it, in the same way a senior engineer has always been accountable for a junior's work. Where AI is part of a system we build for you, for example a workflow that calls a language model, the data flows, the provider and the controls are documented in that engagement's design and Statement of Work and agreed with you before anything goes live.

Tools and your information

We use a small, deliberately chosen set of AI platforms, all held on company-managed, business-grade subscriptions under commercial terms that do not permit our content to be used to train the providers' models. Personal and free AI accounts are not used for ITAAAS or client work. There is no shadow AI at ITAAAS, by policy and in practice, and we review our providers' privacy and data-handling terms regularly and before every renewal.

Our work involves ordinary business personal information: names, roles, contact details, what was said in a meeting, and correspondence. We handle it in line with the Privacy Act 1988 and the Australian Privacy Principles: only in our approved tools, only for the purpose it was collected, and only as far as the task needs. Personal circumstances that are not relevant to the work stay out of our records, and AI-drafted meeting notes and CRM updates are reviewed by a consultant before they are saved. 

We tell you when a meeting is being recorded or transcribed, and we don't record if you'd rather we didn't. For work where sensitive data needs to stay offline, we can work within a private, server-based model in our own environment rather than a hosted service.

Where a client has specific requirements about AI use, including a preference that none is used on their work, we follow them and record this in the Statement of Work. Occasionally we will decline an engagement where we don't believe we can do the work properly within these constraints. We'd rather say so up front than compromise on either the work or the data.

Standards we follow

We keep up with government guidance on the safe use of AI and align with it where we can. The two references we work to are the Australian Signals Directorate's guidance on engaging with artificial intelligence and the G7 Hiroshima AI Process guidance for developing advanced AI systems. Our handling of personal information follows the Privacy Act 1988 and the Australian Privacy Principles.

How we tell clients

Going forward, every ITAAAS Statement of Work will include a link to this page. Any client can ask how AI was used on their engagement and we'll tell them plainly. If a client wants a different approach, we agree it up front and write it into the Statement of Work.

Keeping this current

The tools change quickly and so will this statement. We review and update it at least every six months, at each subscription renewal, and whenever our practice changes. If you think we've got something wrong or could do better, tell us. We're open to feedback and we'll act on it. Questions and suggestions go to Patrick Devlin, Managing Partner, at patrick.devlin@ITAAAS.com.au.

In the spirit of the above: this statement started as our own position, was drafted with the help of an AI tool, and was then reviewed and revised by the ITAAAS team before publishing. It DID take us MANY attempts to get it right. Humans came up with the idea and concept, we peer reviewed it, AI helped draft it then we peer reviewed it again, and again, and again, and again. V1.1 took the team around 11 edits and feedback sessions. That's how we work.