Each writing profile is fully independent. Profile A and Profile B do not share a sandbox, do not share version history, and do not share Layer 3. A change to one has no effect on the other.
The independence is at the Layer 3 level. Layers 1 and 2 — the global baseline and the niche template — apply universally across every profile on your account. Layer 3 is the personal voice layer, and that is what differs between profiles. If you write a weekly newsletter in a dry, essayistic register and a technical product blog in a precise, instructional one, those are two different Layer 3 configurations. Two profiles.
Think of each profile as a living document rather than a fixed setting. As your voice evolves, your audience shifts, or your sandbox sessions reveal mismatches between the profile and your actual style, the profile should be updated to reflect that.
How profile selection works in the pipeline
Profile selection happens at the idea card level, before drafting begins. When you kick off the pipeline, you choose which profile to run it against. That selection governs all three pipeline stages — research, outline, and draft.
You can change the profile on an idea card before drafting starts. Once drafting begins, the profile is locked for that piece. If you need a different profile, that decision needs to happen earlier in the process.
Profile names appear directly on the idea card during selection. A name like "Newsletter" or "Technical Blog" is functionally useful — you are making a quick selection in context, not navigating a settings menu. Names that clearly reflect purpose reduce the chance of running the wrong profile against the wrong piece.
Setting up a new profile
Creating a new profile follows the same six-step wizard as your initial profile setup. It is not a shortcut or a duplication of an existing profile — it builds a fresh Layer 3 from the beginning.
After you complete the wizard, the new profile requires its own sandbox tuning session. This is not inherited from any existing profile. Each profile's sandbox is independent, and a new profile starts without any tuning history. You will need to run the sandbox, review the output, and calibrate the profile before using it on live content.
The sandbox is the mechanism for testing and refining how the profile performs. Version history is the recovery mechanism if an update to a profile degrades output — you can roll back to an earlier version without affecting any other profile on your account.
Plan for the tuning session before you expect to use a new profile in production. A profile that has not been through the sandbox is a profile that has not been tested.
Using Multiple Profiles
The decision rule is straightforward: would a draft generated for one context need substantial editing before it fits another?
If the answer is yes — different audiences, different registers, different structural conventions — then separate profiles are the right tool. If the answer is no, or close to no, then a second profile is not adding value. It is adding maintenance overhead.
One well-tuned profile outperforms two shallow ones. If you are considering a second profile because your first profile feels off, the better intervention is usually more sandbox work on the profile you have, not a new one. Depth in a single profile produces better output than breadth across multiple underdeveloped ones.
A second profile is warranted when the difference in your writing is real enough that a single profile cannot represent both without compromise. That threshold is higher than it might initially seem.