For accountants, bookkeepers and other practices holding a confidentiality obligation. What happens to client information once it is pasted into an AI tool — where it goes, what a provider retains, and what the practice should have written down.
Ask about a sessionCognivren runs the session on site for the whole team at once, rather than sending one person away to a webinar. It takes about two hours, is built around the tools the practice already has in front of it, and is scoped in a short conversation beforehand so the examples are the practice's own.
Most guidance on AI risk stops at the policy layer: who is allowed to use which tool, and who signs that off. The question underneath it is more concrete. When someone types a client's figures, correspondence or working papers into an AI tool, where does that text go, who is able to read it, how long is it held, and can it affect what the tool produces for someone else later?
The answer is not the same for every tool, and not the same for every account on the same tool. A personal account on a free plan, a paid business plan, and an AI feature built into software the practice already licenses can sit in genuinely different positions on retention and on whether submitted content may be used to improve the underlying model. The session works through that distinction on the tools the practice actually uses, and separates what a provider commits to in its terms from what is only being assumed.
It also covers the exposures that are easy to miss. Content does not have to be stored long term to have left the practice, and material that looks harmless on its own can become identifying once it sits alongside the rest of a file. None of this leads to the conclusion that AI tools cannot be used. It leads to knowing which tasks carry little risk, which do not, and where the practice needs a standing rule rather than a judgement call made in the moment.
Built for practices holding client material they are obliged to protect, and for businesses in the same position without the professional title.
The material the session works through, on the practice's own tools rather than in the abstract.
The session is built to produce something the practice keeps, rather than notes that go in a drawer.
Ten years building and running production AI and machine-learning systems, a Masters in Computer Science, and delivery experience across ecommerce, industrial equipment and professional services. That work is the data handling underneath these systems, the way they are deployed, and what they hold on to while they run. The session comes out of it.
That is the reason it can answer the systems question and not only the governance one. Most available material on AI risk is written at the level of controls and sign-off, which matters, but it leaves the layer underneath unaddressed: where the data physically goes, what is kept, and what can be recovered or inferred from it afterwards. A session on client confidentiality is only as good as its answer to that.
Cognivren is the trading name of Revodien Pty Ltd. The session is training rather than legal advice, and it does not replace a practice's own judgement about its obligations or advice from its own advisers.
Continuing professional development is recorded by the member, not certified by a provider. Professional bodies do not confer a formal status on individual activities, and Cognivren does not claim one.
What the requirements generally turn on is whether an activity is relevant to the member's work, and whether the provider holds suitable qualifications or practical experience in the subject. Confidentiality also sits among the fundamental principles of the codes of ethics that apply to accounting practices, which is the ground this session stands on.
The outline above is deliberately specific so that a member, or a practice manager, can hold it against their own body's requirements and decide for themselves. Cognivren does not make that assessment on anyone's behalf. An outline of the material and a record of attendance are provided either way.
The same tools and the same workflow material, without the confidentiality framing, are covered on general AI training for teams.
Sessions are scoped before they are booked — what the practice uses, what it is worried about, and what it wants written down by the end. A short conversation is enough to work out whether it is worth doing.