AI Policy

AI proposes; it never sends. Nothing generated by AI reaches your client until you have reviewed and approved it. Section 4 sets out exactly what each AI provider may do with the content you send it — including where their rules differ.

AI in Waco3 drafts things for you: proposal sections, follow-up messages, survey questions, summaries, and company research. It works from the context you give it — your brief, your documents, your client’s engagement. Every output lands in your workspace as a draft. There is no mode in which AI contacts a client, sends a document, or commits you to anything on its own.

Every AI feature in Waco3 is something you press a button to use. Nothing analyses your content unless you ask it to — there is no background scoring, no silent enrichment, no passive classification of what you write. Some requests (like client research) finish in the background after you start them, but they always start with you.

Waco3 does not build or train artificial-intelligence models, and we never use your content to do so. We send your content to third-party AI providers to produce the output you asked for. Your prompts, your documents, your recordings and your outputs go to no AI provider other than the ones named below.

Which provider handles a given request is chosen automatically, by cost and availability, and can differ from one request to the next. We cannot tell you in advance which one it will be — which is why we tell you the whole set instead of naming one.

The providers named in section 4 act as our service providers under our Privacy Policy, which lists every provider we use and explains how international transfers are handled. Content sent to DeepSeek is processed in China, which is outside the EEA and the UK and is not covered by a European Commission adequacy decision — if that matters for your work or your clients’ data, please take it into account before using AI features.

We may change the models or providers behind AI features at any time, provided the commitments on this page continue to hold — no training of Waco3 models on your content, accurate disclosure of what each provider may do, and review-before-send. If a change would materially reduce these commitments, we will tell you at least 30 days before it takes effect, and update this page and the Privacy Policy.

As between you and Waco3, you own what you put in and what the AI produces for you, and we assign to you any rights we may have in that output. Two honest caveats: AI output may not be unique — similar prompts can produce similar results for other users — and purely machine-generated content may have limited copyright protection in some jurisdictions.

AI output is generated by statistical models. It can be wrong, outdated, or confidently misleading — especially about numbers, legal terms, and facts about companies. You are responsible for reviewing every output for accuracy, tone, and suitability before it goes anywhere. AI output is not legal, tax, accounting, or financial advice, and we give no warranty or indemnity for it.

AI features are for your own client work inside Waco3. As set out in section 12 of our Terms of Service, you agree not to use them in an automated or scripted way or to circumvent credit limits; to build or train a competing AI product or extract the underlying models; to resell or proxy them; to harvest data about individuals or contact people who have not dealt with you; or to make automated decisions with legal or similarly significant effects about a person without human review. AI proposes; a human decides.

AI usage is measured in credits included with your plan; the Terms of Service and the pricing page describe allowances, Boost packs, and expiry. When credits run out, AI features pause — nothing else in your account is affected.

Client research looks a company up in public sources — public web search (Brave Search), public company registries (OpenCorporates), and the company’s own website, which our servers visit and read — and runs the result through an AI provider to build a summary. It can surface the names and public professional details of people listed on those sources. Sources can be wrong or out of date, and a research summary is a starting point, not a verified fact — check it before you rely on it in a proposal. Privacy Policy section 10 explains what we keep and for how long.

Voice input lets you dictate instead of typing. Your device records the audio and sends it to us, and we forward that recording to a speech-to-text provider — Groq, OpenAI or Google — to turn it into text. The recording is held in memory only for as long as the transcription takes, and is never written to our storage — not to a file, not to a database, not to your account. What we keep is the text it produced, saved as your prompt exactly like anything you type. Once it reaches the provider, that provider’s own retention rules apply (see the table in section 4); our promise covers our systems. Voice is always optional: every field that offers the microphone also accepts typing. You are responsible for complying with any recording-consent laws that apply where you and anyone you record are located.

Where a feature involves interacting with an AI system, we label it as AI in the product, and the label appears no later than your first interaction with it. Where our AI generates content, we are implementing machine-readable indications that the content was AI-generated, in line with Article 50 of the EU AI Act.

Drafts in Waco3 are made for your review: once you review, edit, and take responsibility for a draft, it is your content and it goes out under your name — which is exactly the human editorial control the rules are designed to preserve.

This policy is part of our Terms of Service. If we materially change how AI features handle your data, we will notify you by email and in the app at least 30 days before the change takes effect, and update the date at the top of this page. Questions? Email [email protected].