Newsletters, posts, articles in the client's voice
Draft with one model from the voice examples; review with another for anything off-brand. The client approves; you post.
An agency selling AI services lives or dies on context: the client's brand voice, their data, what was promised in the proposal, what was delivered last month, what the client hated. When that context lives in one Project per client, every proposal, report, automation and revision comes out consistent, and any AI you use reads the same brief. This page is how to set an AI agency up that way, from proposal to delivery to renewal.
Free to start with one client Project. More Projects and the latest models when the client list grows.
Who they are, what they sell, their voice with three examples, what they never say, who approves work. Every model reads this before every task.
Brand guide, past campaigns, product sheets, customer lists, the spreadsheet they run the business on. Upload directly or import from Google Drive where enabled.
What was promised, for how much, by when. When the client asks for more, the AI can tell you it is outside scope before you agree to it for free.
Each deliverable as a file, each revision with the client's feedback next to it. Six months in, "make it like the March campaign" is a question the AI can answer.
If you build something for the client, the code or workflow and the samples of what it produces live here, so you can maintain it and report on it.
Generated from the Project: what was delivered, results, what is next. A spreadsheet or a deck, from the same context, in minutes.
Draft with one model from the voice examples; review with another for anything off-brand. The client approves; you post.
Draft replies in the owner's voice, in bulk, with a second model checking tone. A productised service that scales to many small clients.
Paste the client's numbers; get the report, the chart-ready spreadsheet and the three recommendations, in their template.
A quote calculator, a lead-intake form, a reminder system. Built with AI, reviewed by a second model, maintained from the Project.
Sourced research with a model that cites, checked by another, delivered as a document the client can act on.
Every proposal drafted from what worked for this client, in the structure that won last time.
The agency's asset is the set of client Projects. New staff read the Project and are useful on day one. A client who returns after a year finds you still know their voice. When you productise a service, the Project template becomes the onboarding: instructions, file checklist, first deliverable. And because every model reads the same context, you can switch models per task without re-briefing.
On team plans, Projects are shared across the agency with one invoice and seat management, so a colleague can pick up a client without a handover document. See team projects and managing client projects.
A client cannot tell which model wrote their newsletter, but they can tell when it sounds generic. The latest models, with the client's voice examples in the Project, produce work that needs less editing and passes review more often. With several clients, the paid plan's extra Projects and messages are the difference between a workspace and a workaround.
Newer models hold your whole Project in view instead of the last few messages, reason through trade-offs instead of picking the first plausible answer, and are wrong less often and with less confidence. For the questions on this page, that is the difference between advice that sounds right and advice you can act on.
Pro from $20 a month, cancel any time. The free plan stays free.
Add the client's files, three examples of their voice, and the agreed scope. Then start here.
Files and conversations stay in your account, are encrypted, and are never used to train any model. Keep one Project per client so nothing crosses over. See trust and security.
Yes, on team plans, which include shared Projects, chats, prompts and files, seat management and one invoice.
Projects are for your workspace. Deliver the work as files and reports; keep the working context yours unless you choose to add a client seat.
The free plan includes one Project. Paid plans include many more, sized for a growing client list. See plans.
No. Tasks and deadlines stay there. The Project holds the context and produces the work.
One Project, file uploads and the major model families, enough to run one client and see whether the setup works for you.
Instructions, files, scope, and the first deliverable from the same context. See how much editing you stop doing. Free to start.
Also: start an AI agency · managing client projects · team projects · analyse client files
Continue learning
Run the same prompt through ChatGPT, Claude, Gemini and Grok before trusting one answer.
Core featureLet several models draft, challenge and verify each other instead of trusting one answer.
ProjectsKeep files, instructions, code and chats together so every model works from the same context.
FeaturesModels, Projects, collaboration, verification, Studios and images in one workspace.
AgencyNiche, offer, delivery system, first clients, then the product.
ClientsContext, files, revisions and feedback in one place.
FilesContracts, spreadsheets and notes questioned by several models.
TeamsShared context, every model, one invoice.
MarketingTwo channels, a weekly routine, one number.