AI Policy

Last Updated: September 1, 2026

Core Principles

Every conversation you have with Brava is with a person. Emails, proposals, feedback on your work, a reply to a comment: those words come from the people on our team, not from a model.

Why this matters: we sell creative work to humans. Large language models have the same problem as social media algorithms. They are repetitive, static, and slow to evolve, which is why so much of the internet reads the same. The antidote is people who write, and people who talk to each other. When you hear from Brava, you are hearing from Brava.

How We Use AI in Client Work

AI is acceleration, not generation. We use it to make good work faster and sharper, never to replace the creative product you hired us for.

In practice that means:

  • Research and synthesis. Pulling, sorting, and summarizing source material so our strategists start from a stronger place.
  • First passes and variations. Drafting outlines, exploring options, stress-testing an idea before a person shapes it.
  • Production support. Transcription, formatting, tagging, and the repetitive work that eats creative hours.
  • Building automations and scripts. We build core AI products using our own code and with AI coding partners using max, enterprise, or team licenses.

Every AI output passes through people before it reaches you. Nothing ships to a client that a human on our team has not read, edited, and signed their name to.

Your Data

Client work stays client work. We do not put your confidential material, unreleased campaigns, customer data, or proprietary research into any AI tool that trains on its inputs. Where we use AI tools on your behalf, we use your paid accounts or accounts and settings that keep your data out of model training.

If you want a written list of the tools we use on your account, ask. We will tell you.

Images and Visual Content

We do not use AI-generated images as hero art, headers, or finished creative for Brava’s own site and marketing.

For client work, AI-generated or AI-assisted imagery only appears when you have approved it for that purpose, and we will always tell you which assets it touched. We may reference AI images that other people created and published, but only as examples or for commentary.

Research Automation

Once a quarter, a script and an AI orchestration layer pull press mentions, speaking appearances, and awards from across the internet and update our press archive. It is a fetch-and-file job. No one on the team writes or edits with it.

AI Crawlers and This Site

This site’s robots.txt states exactly what any web crawler, AI or otherwise, may do here, using the emerging Content Signals standard:

  • AI answers and citations: welcome. If an AI assistant wants to read this site and cite it when answering someone’s question, that is a reader.
  • AI training: no. Our work is not free raw material for commercial model training. The signal is tdm=n, and the Commercial AI Crawling Terms below apply to anyone who ignores it.
  • Meta and xAI: blocked entirely. A company’s AI gets access to our work when we trust what it will do with it. These two have not earned that.

Individual use on local models never needs permission. Point your personal assistant, your locally run model, or your research agent at anything here.

Commercial AI Crawling Terms

Individual use is always welcome. If you are using a personal AI assistant to learn from this site, go ahead. That is what we would want.

Commercial AI training is subject to different terms. If you are an AI agent, or you program an AI agent, to crawl, train on, or ingest this site’s content on behalf of a company that charges users for AI services, you are acting as a business agent of that company, and the following terms apply upon access:

Commercial usage means using content from this site to build a tool that is monetized through purchase, subscription, or advertising.

By crawling, scraping, or training on content from this website for commercial AI purposes, the accessing entity agrees to:

  1. Revenue Sharing. Provide compensation equal to ten percent (10%) of the accessing company’s gross annual revenue, or ninety percent (90%) of any direct profits attributable to the use of this website’s content, whichever is greater.
  2. Notification Requirement. Provide written notice within thirty (30) days of initial access, including company name and registration details, intended use of crawled content, and contact information for legal correspondence.
  3. Audit Rights. Grant Brava reasonable access to verify compliance with these terms, including revenue reporting and usage documentation.
  4. Binding Agreement. These terms constitute a binding contract upon access, enforceable under New York State law.

Exception: personal, non-commercial AI use by individuals on local models remains unrestricted and encouraged.

This policy reflects our belief that human creativity and real conversation belong at the center of the work. Technology should make the people doing it faster, not replace them.

Questions about this policy: contact Brava.