On this page
- Why the therapist stays in the picture
- What you can set up
- AI support in your approach
- Worksheets and exercises
- Assessments
- A first month with one client
- What runs without you
- Where it stops
- Frequently Asked Questions
- What can a therapist set up for a client between sessions in Citt?
- Who decides when worksheets and follow-up assessments go out?
- What happens if a client writes something that suggests risk?
- Is there evidence that between-session support improves outcomes?
In Citt a therapist can set up four things for a client between sessions: AI support in the approach they choose, worksheets and exercises, assessments, and messages on WhatsApp. This post describes each, what runs without the therapist, and where it stops.
Picture a client who leaves a Thursday session with a thought record to try before next week. By Monday the sheet is folded into a bag, and by Wednesday she remembers it only as something she meant to do. When she comes back, the session starts with why the homework didn't happen.
In a 2016 meta-analysis of 17 studies of cognitive behavioral therapy, covering 2,312 clients, Kazantzis and colleagues found that clients who completed more of their homework, and did it with more skill, had better outcomes.1 The relationship is an association, so it does not establish that homework causes the improvement.
Why the therapist stays in the picture
In a 2012 meta-analysis of 19 randomized trials of computer-based treatments for depression, Richards and Richardson found a moderate effect overall, and the programs that included human support had better outcomes and kept more of their participants.2 Mohr and colleagues proposed a model for why. Their supportive accountability model holds that human support increases adherence because the person feels accountable to someone they see as trustworthy, well-meaning and expert.3 It is a model of human support with hypotheses still to test, and it says nothing direct about AI.
My view is that software between sessions should extend a relationship the client already has, with a therapist they know is paying attention, and these findings are consistent with that. In Citt the therapist sets up each piece, and clients are told in the product that their therapist can see their conversations with the AI, though not always right away.
What you can set up
AI support in your approach
You set up AI support for each client and choose the approach it follows: CBT, DBT skills, ACT, mindfulness-based, or a focus area such as sleep, stress or social anxiety. You can also write your own instructions. The client can then talk to it between sessions, in the app or on WhatsApp. There is no cap on how long or how often. Choosing the approach the AI support follows explains what that choice changes.
The AI's instructions steer it away from heavy reassurance and from encouraging avoidance, and toward bringing recurring themes back to you. You can read a client's conversations with the AI whenever you choose.
Worksheets and exercises
Citt has a library of worksheets and exercises drawn from CBT, ACT, DBT, exposure and response prevention, and compassion-focused work, along with resources on neurodiversity. You can send one to a client by WhatsApp, by email or as a link, so the worksheet from the opening can reach her phone the evening after the session, in a channel she already reads.
Assessments
When a client starts, a baseline set of questionnaires goes out: PHQ-9, GAD-7, the Perceived Stress Scale, WHO-5 and the Multidimensional Scale of Perceived Social Support. After that, Citt tells you when a follow-up is due, at 4, 8 and 12 weeks and at 6 months, and you decide whether to send it. Clients do not see their scores. You see the scores and the change from baseline. What each of these scales adds to the PHQ-9 and GAD-7 goes through them one by one.
WhatsApp is the default channel for the messages a client receives outside the app. It runs on Meta's WhatsApp Business API. A client can talk to the AI there, receive check-ins and the resources you send, and get appointment reminders. Messages on WhatsApp get the same safety screening as the app and are stored in the client record. Citt has no business associate agreement with Meta, so use the channel with the client's consent and in line with your own professional and legal obligations.
A first month with one client
As an illustration, this is how the pieces could fit together for the client from the opening.
- When she starts, the baseline questionnaires go out to her.
- You set up her AI support and choose CBT, because that is the work you are doing together.
- After Thursday's session you send her the thought record from the library, on WhatsApp.
- On Monday evening she talks one entry through with the AI.
- Before the next session you read the summary, and the conversation itself if you want the detail.
- At four weeks Citt tells you a follow-up is due, and you decide whether to send it.
Only steps 2, 3 and 6 need a decision from you.
What runs without you
Once a client is set up, these parts run without your input:
- The baseline set of assessments goes out when the client starts.
- Every message a client sends to the AI is screened for risk before the AI replies, and the screening cannot be switched off.
- Citt tells you when a follow-up assessment is due.
- Clients get appointment reminders on WhatsApp.
- Before each session, Citt summarizes check-in trends, conversation themes and assessment changes since the last one.
The screening has two levels. A message treated as a crisis shows the client crisis guidance and alerts you, while less explicit language is only recorded. Citt is not an emergency service. How crisis screening works between sessions sets out the details.
Where it stops
After the baseline, worksheets and follow-up assessments are yours to send. The AI follows the approach and instructions you set.
There is no evidence yet that Citt improves clinical outcomes. The research in this post is about homework in CBT and about human-supported digital programs, and it does not show that these particular tools help. Our Trust page sets out what we have evaluated so far.
Footnotes
-
Kazantzis, N., Whittington, C., Zelencich, L., Kyrios, M., Norton, P. J., & Hofmann, S. G. (2016). Quantity and quality of homework compliance: A meta-analysis of relations with outcome in cognitive behavior therapy. Behavior Therapy, 47(5), 755-772. Seventeen studies, 2,312 clients. Compliance was related to outcome at posttreatment and at follow-up; only three studies measured the quality of homework, so that estimate is wide. ↩
-
Richards, D., & Richardson, T. (2012). Computer-based psychological treatments for depression: A systematic review and meta-analysis. Clinical Psychology Review, 32(4), 329-342. A review of 40 studies with a meta-analysis of 19 randomized trials (pooled post-treatment d = 0.56). The programs studied predate generative AI. ↩
-
Mohr, D. C., Cuijpers, P., & Lehman, K. (2011). Supportive accountability: A model for providing human support to enhance adherence to eHealth interventions. Journal of Medical Internet Research, 13(1), e30. A theoretical model drawn mainly from organizational psychology, motivation theory and research on computer-mediated communication. It concerns human coaches, and its components are presented as hypotheses to test. ↩
Frequently Asked Questions
What can a therapist set up for a client between sessions in Citt?
You can set up AI support that follows the approach you choose, send worksheets and exercises by WhatsApp, email or link, and send follow-up assessments when Citt tells you they are due. Clients can also receive check-ins and appointment reminders on WhatsApp, the default channel for messages outside the app.
Who decides when worksheets and follow-up assessments go out?
You do. A baseline set of assessments goes out when a client starts, Citt tells you when a follow-up is due, and you decide whether to send it and which worksheets to send.
What happens if a client writes something that suggests risk?
Every message is screened before the AI replies. When a message is treated as a crisis, the client is shown crisis guidance for their region, a care record is created, and you get an alert in the product and an email; less explicit language is recorded without an alert. Citt is not an emergency service.
Is there evidence that between-session support improves outcomes?
Homework completion in CBT is associated with better outcomes, and a 2012 meta-analysis found that computer-based treatments for depression did better when they included human support. There is no evidence yet that Citt's tools improve clinical outcomes.
Stop rewriting notes from scratch
Citt.ai drafts session notes under your review, keeps between-session context in one place, and stays therapist-led. Start free. No card required.
Related Articles
- The 167 Hours Between Sessions: A Client's Guide to AI Support
If your therapist has offered you AI support between sessions: what it is, what it cannot do, who can read it, and what happens in a crisis.
- Where a Therapist's Admin Hours Go, and What Software Can Take
Which parts of a therapist's work outside sessions software can take, which it can draft for review, and which stay with the therapist.
- How Citt.ai Screens Client Messages for Risk, and Where It Stops
What happens when a client writes something alarming to AI support between sessions: how Citt.ai screens messages, who is alerted, and what it misses.