AI & Data Practices

Effective August 25, 2026

We sell AI systems, so we owe you a plain description of how we use AI and how we handle the material you trust us with. These are commitments, not marketing.

AI on this website

There is no chatbot, AI assistant, or automated agent on this website. Nothing here talks back, and nothing here is analyzing your visit. When you email us, a person reads it.

AI in the systems we build for clients

Disclosure

Voice and chat agents we build identify themselves as AI assistants at the start of an interaction. We do not build systems designed to be mistaken for a human being. Where a client asks us to, we will decline.

Human oversight

Systems assist; people decide. Anything that goes to a customer as a commitment, anything that becomes a contract, and any public claim gets human review before it leaves. Drafting can be automated. Responsibility cannot.

Grounding and refusal

We build agents to answer from client-approved source material and to escalate to a person when a question falls outside it. A system that invents a plausible answer to seem helpful is worse than one that takes a message, and we tune for the latter.

Access

Systems connect only to what a client explicitly approves, scoped to what the capability actually requires, and access can be revoked at any time.

How we handle client material

We do not use client material to train AI models, and we configure the model providers we use to exclude client content from training.
We do not sell client or end-customer data, and we do not share it for advertising.
Subprocessors are disclosed. Delivering these systems means using third-party model providers, hosting, and communications platforms. Which ones apply to a given engagement is identified in the client agreement.
Data-handling terms are contractual. Where a client's own legal obligations require a data processing agreement, we sign one.
Retention is agreed, not assumed. What is stored, for how long, and what happens at the end of an engagement is written into the scope.

Voice agents, calls, and messaging

Telephone and messaging systems are heavily regulated, and the rules turn on things like who initiated the contact, what consent exists, and which states are involved. Our standing practice when we build these systems:

The agent discloses that it is an AI assistant.
Where calls are recorded or transcribed, we configure an audible notice at the start of the call and treat all-party consent as the default, rather than relying on Texas's one-party consent rule when the other party could be anywhere.
Outbound calling and texting is configured around the consent the client actually holds, with working opt-out handling and honest caller identification.
We will not build a system whose purpose is to evade these rules.

Compliance for a client's own outbound program depends on that client's consent records and practices, which we do not control. We raise the issues and build to the safer default; we are not your lawyers.

What we will not build

Systems designed to deceive people about whether they are talking to a machine
Systems intended to evade consent, opt-out, or do-not-call requirements
Systems that make claims about a business we have no basis to believe are true

Questions

Procurement, security, or data-handling questions: hello@vedennius.com. Privacy requests about this website are handled under our Privacy Policy.