We use AI tools to deliver our work. If you are paying us, you should be able to see where we use them, where we do not, and what keeps your information safe.

Read this alongside our Privacy Policy and our Terms and Conditions. If you are a client, your proposal and your services agreement are the agreement between us, and they win wherever they say something different. This page explains the detail.

Our position

AI-enabled. Human-accountable.

AI lets one person take on more than one person otherwise could. It does not replace professional judgement, and it does not carry responsibility for the work. Everything made with the help of an AI tool is made under our direction, and it is part of a human-led professional service.

What this policy covers

Generative AI, machine learning tools, AI features built into software we already use, and chatbots. It covers the AI tools we use and control while doing your work.

It does not cover AI features inside a product you use directly, unless we have said so in writing. It also does not cover a system we have built for you once it is running in your business. Your agreement covers that.

Where we use AI

Research, and summarising documents, policy and background material
Structuring and drafting documents, plans, processes and written outputs
Analysing information you have given us
Editing, formatting and quality review
Designing, building and testing the systems you have asked for
Our own admin and internal workflow

Where we do not put your information

We do not enter the following into AI tools:

Your confidential information or personal information, without your express written consent
Information that identifies a person: your staff, your customers, your contractors, your suppliers
Material you have told us is culturally sensitive, or that is subject to a community protocol
Sensitive information, as that term is defined in the Privacy Act 1988 (Cth): health, racial or ethnic origin, political opinion, religious belief, union membership, sexual orientation, criminal record, biometric data

Where it is reasonably practicable, we ask you to take names and identifying details out before you send information to us. Where we cannot avoid handling identifying information, we remove or mask those details before anything goes into an AI tool, unless you have told us in writing that we can do otherwise.

Decisions about people

We do not use AI to make a decision about a person, or to substantially help make one. That includes decisions about employment, pay, promotion, discipline, eligibility, access to a service, credit or risk.

If we build you a system that touches a decision like that, a person in your business makes the decision and the system supports them. We say so plainly in what we build and in what we hand over.

Evidence and claims

We do not use AI to produce evidence, citations, statistics, quotes or claims that we have not checked ourselves against the original source.

Where something we give you contains a factual claim, a number, a citation, a quotation or a calculation, it has been checked against the source material before it reached you.

Human oversight

Everything we deliver is reviewed, checked, edited and approved by Steph before it reaches you. AI is not used to produce a finished deliverable without that review.

We keep human responsibility for working out what information means, drawing conclusions, making recommendations, reporting on results, and noticing where something could cause harm you have not thought of.

AI can be wrong, and sometimes it is wrong confidently. We stay fully responsible for what we hand you and for the advice we give you about it.

Fairness

We are committed to using AI fairly and without discrimination, and we take reasonable steps to notice and reduce bias in the tools we use and in the information behind them.

Where AI has drafted something that describes a group of people, we check it against your own words rather than accepting whatever the tool produced by default, and we read it again for stereotyping and for assumptions nobody actually made.

What we will never use AI for

Anything unlawful
Discriminating against a person on the basis of a protected attribute
Producing or spreading misleading information
Infringing anyone else’s intellectual property, or copying someone else’s work
Creating deepfakes or manipulated media
Facial recognition, biometric identification or tracking people
Making a decision that affects a person without a human involved in a meaningful way

Talking to an AI, not a person

If you or your customers are ever dealing with a chatbot or an automated assistant rather than with Steph, we say so at the point it happens. We do not pass off an AI as a person.

Your choices

Ask us at any time whether AI was used on a piece of your work, and how.
Ask us not to use AI tools on your project, or on one specific piece of work. You do not have to give a reason. If it changes how long the work takes, we will tell you before we start and agree any adjustment with you first.
If your business has its own AI use policy, tell us and we will work to it.
If something we produce is going to a funder, an insurer, a regulator or an auditor who requires AI use to be disclosed, tell us what they require and we will meet it.

The AI tools we use

This is the register. It says what each tool is used for, whether your information goes into it, and what stops that information being used to train somebody’s model.

Tools that handle client material

Claude, by Anthropic. Research, drafting, structuring, analysis, building and admin. This is the main tool we work in.
Training: switched off in the account settings. Checked on 3 September 2026, and checked again whenever we change plan or open a new account.

Notion AI. Searching and summarising inside our own workspace, where client notes and project records are kept.
Training: Notion states that it and its AI subprocessors do not use customer data to train any models. This is contractual and applies to every plan, so there is no setting to get wrong.

AI tools we use only on our own material

Your information does not go into these. They are used on our own marketing and teaching content.

Whimsical, for diagrams. Whimsical states that it does not use customer data, and does not let others use it, to train the models behind its AI features.
Tella, for recording and editing our own videos.
ElevenLabs, for voice in our own video work.

Some of these tools store information outside Australia, including in the United States. We choose providers that publish a privacy policy and commit to protecting what they hold. Our Privacy Policy covers how we handle information held overseas.

If we start using a new AI tool on client work, it goes on this list before it touches anything of yours.

Your ownership is not affected

Using AI in our process does not change what you own. The finished work is yours, on the terms in your agreement with us. Where something was made with AI help and it is not clear that copyright applies to it, you get a permanent, unrestricted licence to use it in your business instead. Either way, you can use it, change it and keep it.

Our own methods, templates, structures, prompts and teaching material stay ours, as your agreement sets out.

Who else touches your work

Steph Does Systems is a one-person business, and Steph does the work. If we ever bring someone in to help on your project, we will tell you first, and they are bound in writing to this policy before they touch anything of yours.

If you think we have got this wrong

Email hello@stephdoessystems.com. We will look into it, tell you what we find in writing, and tell you what we are doing about it.

Keeping this current

We review this policy at least once a year, and we update it when the tools change, when a client needs something different, or when the rules change. The date at the top shows when this version was published.

Contact

Stephanie Bartram, trading as Steph Does Systems
ABN 93 243 679 794
South West Rocks NSW 2431, Australia
hello@stephdoessystems.com

AI Use Policy