AI-use disclosure

The short version

AI helps us draft content and powers features like the readiness quiz and DIAGNOSE. A named human reviews and verifies the things that matter before they reach you. We don’t train AI models on your identifiable data without your explicit consent, and our AI providers don’t train their models on the data we send them. This is the practical side of our Amplify-Don’t-Replace Pledge: AI should amplify people, not be used against them.

Where we use AI

• Drafting content — guides, The Signal, templates and explainers are often drafted with AI, then edited and fact-checked by a person. Every published piece carries a “drafted with AI, verified by a human” badge naming the reviewer and the date.
• The readiness quiz and DIAGNOSE — AI interprets the answers you give about your business to produce a readiness score, a plan and recommendations tailored to your industry, size and constraints.
• Search and recommendations — we use AI-generated embeddings to power relevant search results and to suggest the next useful thing to read.
• Safety checks — AI helps us flag inputs that need a human (for example, regulated or high-stakes situations) and redact obvious personal or secret data before it’s stored.

Where a human is always in the loop

AI drafts; humans decide. We never publish AI output as verified without a person checking it against sources. For paid DIAGNOSE plans, a human reviews the plan before it’s treated as final. Where a situation genuinely needs an expert — legal, regulatory, safety or compliance — we say so rather than letting a model pretend otherwise.

Which AI we use, and how

We use third-party AI providers to run these features: Anthropic (Claude) for language tasks and Voyage AI for the embeddings behind search. We send them only the data needed to perform the task. Under our agreements with these providers, the data we send through their APIs is not used to train their models. See the full subprocessor list for who processes what and where.

What we never do

• We don’t train AI models on your identifiable data without your explicit, opt-in consent.
• We don’t let AI make a final, consequential decision about you without a human able to review it.
• We don’t sell your data, and we don’t use it for advertising.
• We don’t use AI to recommend replacing your people — our guardrails actively block “use AI to make layoffs” goals, in line with the Pledge.

Improving our AI — only with consent

Separately and explicitly, you can choose to let us learn from your inputs to improve our AI features. This AI-training consent is off by default and never assumed. If you opt in, we only ever learn from anonymised, aggregated patterns — never your raw identifiable data — and you can withdraw consent at any time. Records without this consent are excluded from any model-improvement process by default. Manage this choice on the cookies & consent page.

AI can be wrong

Even verified, AI-assisted content is guidance, not professional, legal, financial or regulatory advice for your specific situation. We work hard to be accurate and to cite our sources, but you stay responsible for decisions you make for your business. If you spot something wrong, tell us at hello@vocabotics.com and a real person will look into it.

Changes to this disclosure

If we change how we use AI in a way that matters to you, we’ll update the date above and flag it. We won’t quietly broaden it.