Reflective thinking chat
A reflective thinking chat scoped to one client at a time.
Every client folder on the Professional tier has a thinking-out-loud chat that draws on that one client's plan and recent drafts only, so you can compare frameworks, think through sequencing, or sit with a dilemma, without receiving directive advice and without anything being cross-referenced against another client's material.
A space to think out loud, anchored to one client's work.
Some clinical questions are too small for supervision and too big to swallow on your own, like whether to bring in a compassion-focused thread alongside the CBT consolidation, whether the parts work that is opening up has the right protector frame, or whether a recent shift in social-withdrawal evidence is enough to update the formulation now rather than at the next plan-and-review. The reflective chat is the place to articulate questions like those, with a thinking partner that has read the same notes you have, that is anchored to that one client, and that replies in the voice of peer consultation rather than directive advice.
Boundaries the chat refuses to cross.
The reflective chat refuses certain moves by design, so the clinical work stays yours from the question to the answer.
- 01
Will not diagnose.
The chat will not offer a diagnostic conclusion or differential, because diagnosis is your professional act, made with the assessment material in front of you rather than produced as a chat reply.
- 02
Will not recommend treatment.
It will not tell you which intervention to use, which protocol to follow, or which medication to consider, but it can help you articulate the options you are weighing while leaving the choice with you.
- 03
Will not predict risk.
Risk is your formulation, so the chat will not generate a risk score, predict an outcome, or substitute itself for your clinical judgement.
- 04
Will not cross client folders.
What lives in one client's folder stays there, with the chat anchored to that client's plan and recent drafts only, so it cannot draw analogies to other people you see.
The same anonymisation, UK hosting, and contractual posture as everything else in Cogent.
The reflective chat uses the same in-browser anonymisation flow as the rest of Cogent, so identifying details are masked before any message reaches the AI, with the mapping back staying on your device and the chat history encrypted under your key, stored under the client folder, and never used to train any model. You can clear a chat, archive a folder, or delete it entirely from Settings, under the same one-click rules as everything else in the product.
A separate chat for the questions that are not about one client.
Some clinical questions are not about a specific client at all, like how schema therapy handles polarised modes when the schemas are entwined, what the recent neurodevelopmental literature says about combined-presentation work, or how a peer supervisor might frame a difficult conversation with a regulator. The general chat sits outside the client folders for exactly that kind of question, draws on the most capable model available for clinical-knowledge questions, and answers in the same peer-consultation voice as the per-client chat, while declining the same things it declines there, including diagnoses, risk predictions, and drafting work that belongs in the workspace. Because no client material is loaded automatically and there is no folder to mask against, the general chat is the one place in Cogent where the clinician is responsible for keeping identifying details out of the conversation by hand.
Try it on this week's work.
The free tier covers ten generations a month with no card to start and no time limit on the trial, so you only upgrade when Cogent has earned at least the cost of a session in time saved.
Start free