Ventiq designs, builds and hands over AI workflows, AI chatbots and websites. Every engagement is defined in a written proposal that sets out the specific deliverables, the systems involved, the agreed timeline and the fixed fee. That proposal is the authoritative description of the work.

Anything not expressly listed in the proposal falls outside the engagement. This includes ongoing operation of the system, staff training beyond any session specified, subscription and usage fees payable to third-party providers, and changes requested after the deliverable has been accepted. We are happy to quote separately for any of these.

On completion we transfer the deliverable to you together with the accounts, credentials and written documentation required to operate it. From the moment of handover the system is under your control, hosted in your accounts, and operated by you or on your behalf.

Handover is the point at which operational responsibility passes to the client. We remain accountable for the system behaving as specified at the time it was delivered. We are not able to be accountable for how it is subsequently configured, extended, retrained, connected to other systems, or used.

AI systems are tools. Like any tool, their effect depends on how they are operated, what they are pointed at, and the judgement of the people using them. Once a system is in your hands, you are responsible for the purposes to which it is put and for the outcomes of those purposes.

Ventiq accepts no liability for misuse of any system we have delivered. Misuse includes, without limitation: using a system for a purpose materially different from the one specified in the proposal; supplying it with data you are not entitled to use; disabling, bypassing or ignoring safeguards, review steps or human approval points that formed part of the delivered design; publishing AI-generated output without the review the design assumes; and using the system in a manner that breaches applicable law or the terms of any third-party provider.

Where we have built a human review step into a workflow, that step is part of the deliverable and not an optional extra. Removing it changes the risk profile of the system, and responsibility for that change rests with the party who makes it.

We build strictly within the access a client grants us and according to the instructions a client gives us. We do not seek broader permissions than the work requires, and we ask clients to grant the minimum access necessary.

We are responsible for building competently and securely within that access. We cannot be responsible for consequences arising from access we were not given, systems we were not shown, constraints we were not told about, or instructions that were incomplete or inaccurate at the time they were given. If a requirement is not disclosed to us, it cannot be built for.

AI models are probabilistic. They can produce output that is inaccurate, incomplete, outdated or unsuitable for a particular situation, and they can do so while appearing confident. This is an inherent property of the technology, not a defect introduced by us, and it cannot be engineered away entirely.

We mitigate it: we ground systems in your own documents and data where possible, we constrain what they are permitted to do, we build in error alerting, and we design review points where the cost of a wrong answer is meaningful. We do not warrant that any AI system will be free from error, and no system we deliver should be treated as a substitute for professional judgement.

Systems that produce legal, medical, financial, tax, employment or safety-related output must be reviewed by an appropriately qualified person before that output is relied upon or acted on.

We treat client data as confidential and use it solely to deliver the engagement. Wherever practical we build inside your existing accounts so that your data does not move to a new location. Where an AI provider must process data, we select business-tier services that do not train their models on customer content, and we tell you which providers are involved and what is sent to them.

You remain the controller of your data and of your customers' data. It is your responsibility to ensure you hold the necessary rights and consents for the data supplied to us or processed by a delivered system, and to meet your obligations under applicable data protection law — including POPIA in South Africa, the GDPR in the European Union, and equivalent legislation elsewhere.

Delivered systems commonly depend on third-party platforms — AI model providers, hosting, messaging services, CRM and accounting tools, and payment processors. These operate under their own terms, pricing and availability commitments, and their fees are payable by you directly.

We are not responsible for outages, pricing changes, policy changes, feature removals or discontinuation by any third-party provider. Where such a change affects a system we built, we will advise on the options and quote for any remediation required.

Every project includes one month of tuning after launch, covering corrections and adjustments to the delivered scope. It does not cover new features, new integrations, or changes arising from a shift in how you run your business.

Beyond that month, a system is supported only if you hold an active care plan. Without one, we have no monitoring in place and no obligation to detect or resolve issues. Software that is connected to changing external services requires maintenance; a system left entirely unmaintained will eventually drift out of correct operation.

To the fullest extent permitted by law, Ventiq's total aggregate liability arising out of or in connection with an engagement is limited to the fees actually paid by the client for that engagement.

We are not liable for indirect or consequential loss, including loss of profit, loss of revenue, loss of anticipated savings, loss of business opportunity, loss of goodwill, or loss or corruption of data, whether or not such loss was foreseeable.

Nothing in this policy excludes or limits liability that cannot lawfully be excluded or limited, including liability for death or personal injury caused by negligence, or for fraud or fraudulent misrepresentation.

We decline engagements intended to deceive, impersonate a real person or organisation, generate misleading claims or reviews, evade legal or regulatory obligations, conduct surveillance without the knowledge of those affected, or make consequential decisions about individuals — such as hiring, lending, insurance or eligibility — without meaningful human review.

If a delivered system is later repurposed for any of these ends, that use is a breach of these terms, is undertaken solely at the client's risk, and entitles us to withdraw support.