Most therapy operates on a surprisingly thin information layer. You talk, your therapist listens, takes a few notes, and carries what they can remember into the next session. Important threads get lost. Progress is hard to measure. And the hour you spend in session each week has to carry the entire weight of change.
I’ve built a different kind of practice — one that gives both of us dramatically more to work with.
The Arc of Our Work Together
Weeks 1–2 · Assessment
A 90-minute deep-dive conversation covering your history, your current challenges, and what you actually want to change. You'll also complete a set of validated clinical questionnaires. Everything you share is captured, transcribed, and synthesized — nothing falls through the cracks.
Week 3 · Your Assessment Report
You receive a comprehensive written report: a clinical formulation connecting the dots across everything we've discussed, concrete goals in your own language, and a phased treatment roadmap. This isn't a form letter — it's a detailed, personalized document that becomes the foundation for everything that follows.
Weeks 3–8 · Weekly Therapy + Session Summaries
We meet weekly. After each session, you receive a written summary — not bullet-point clinical notes, but a real narrative capturing what we explored, the shifts that happened, and specific action steps for the week ahead. Therapy doesn't stop when the session ends; the summary keeps the work alive between meetings.
Week 8 · First Progress Report & Check-In
A formal progress report tracking your movement against the goals we set: themes across sessions, questionnaire changes over time, your perceived progress and mine. This is also where we decide on cadence going forward — some clients continue weekly; others shift to biweekly as the foundation solidifies and longer-term work begins.
Ongoing · The Cycle Continues
Weekly or biweekly sessions — whichever suits where you are — each followed by a summary. Progress reports every 6–8 weeks. A living, accumulating record of your growth that makes patterns visible and keeps us accountable to real change.
What Makes This Possible
I use AI tools — specifically, secure enterprise services from Google and Microsoft, accessed programmatically under healthcare-grade agreements — to process session transcripts and draft documentation. This is a core part of how I practise, and it’s what allows me to offer a level of rigour and continuity that traditional therapy simply can’t match.
Here’s what that means in concrete terms:
- A comprehensive assessment report synthesized from our intake conversation, your questionnaires, and my clinical notes — far more thorough than what any clinician could draft by hand in a reasonable timeframe.
- A session summary after every meeting capturing the key themes, the shifts that happened, and specific action steps — written in a warm, conversational tone and sent directly to you.
- Progress reports every 6–8 weeks tracking your stated goals, session themes over time, questionnaire changes, and both your perceived progress and mine.
Without these tools, I’d have two choices: spend hours writing each of these documents myself (unsustainable), or not offer them at all (the industry default). Instead, the AI handles the drafting and documentation while I focus on what matters — the clinical thinking, the interpretation, the decisions about where we go next. That part is entirely mine.
The result is a practice where nothing gets lost, progress is visible, and the work between sessions matters as much as what happens inside them.
Inside the Assessment Report
Below are excerpts from a sample assessment report for a fictional client, Marcus — a 34-year-old software developer navigating anxiety, perfectionism, and the aftermath of a painful relationship. This is the kind of document you’ll receive after our first few sessions.
Presenting Concerns & Goals
Marcus identified three primary goals for therapy:
- Reduce the grip of anxious overthinking — particularly the 2–3 a.m. spiral where he replays conversations and rehearses worst-case scenarios. Success would mean falling back asleep within 20 minutes on most nights rather than losing two or three hours.
- Build a relationship with imperfection — he wants to ship work, share opinions, and make decisions without the paralysing internal audit that currently delays everything. Concretely: submitting code for review without rewriting it a fourth time, and voicing disagreement in meetings instead of composing the "perfect" rebuttal after the fact.
- Understand what happened in his last relationship — not to win Sarah back, but to stop carrying the conviction that his emotional unavailability is a fixed character flaw. He wants to know whether genuine intimacy is something he can learn or whether "there's just something missing in me."
Clinical Formulation
Marcus's anxiety and perfectionism are not separate problems — they are two expressions of the same core belief: if people see the unedited version of me, they will leave. This belief was shaped early. His father's emotional withdrawal wasn't dramatic or abusive; it was simply an absence that taught Marcus a quiet lesson — closeness is conditional on performance, and love is something you earn by being useful, competent, and low-maintenance.
The overthinking at 2 a.m., the compulsive code rewrites, the inability to be emotionally present with Sarah — all of these are strategies that once made perfect sense. When your earliest experience of attachment teaches you that the safest version of yourself is the most polished one, then vulnerability doesn't just feel uncomfortable; it feels genuinely dangerous.
Treatment Roadmap
Phase 1 (Weeks 1–4) — Mapping the machinery. We build a detailed, real-time picture of when and how Marcus's anxiety operates: the triggers, the bodily sensations, the cognitive patterns, and the avoidance behaviours that follow. The goal is not yet to change anything, but to make the invisible machinery visible.
Phase 2 (Weeks 5–10) — Experimentation. Marcus begins running small, structured experiments that challenge the core belief. Submitting "imperfect" code. Sharing an unpolished opinion. Sitting with the discomfort of not having the ideal response. We track what actually happens versus what the anxiety predicted.
Phase 3 (Weeks 11+) — Relationship repair. With a clearer understanding of his patterns, we turn to the interpersonal dimension — exploring what emotional availability actually looks like in practice and building Marcus's capacity to tolerate the vulnerability that genuine closeness requires.
What a Weekly Summary Looks Like
After every session, you receive something like this. It’s not clinical shorthand — it’s a real account of what happened, written directly to you.
Session Summary — Marcus D. (Fictional Client)
Session Focus: The Cost of the Fourth Rewrite and What Happened When Marcus Stopped
We started with the experiment from last week — submitting your code after the second review instead of the usual fourth. You did it, and what you described was fascinating: the anxiety spiked hard for about twenty minutes after you hit submit, then it just… dissolved. Your team lead approved the PR with minor comments. No catastrophe. No judgment. Just a normal Tuesday.
But the more important thing was what you noticed about yourself in those twenty minutes. You said you felt physically lighter the moment you clicked the button — like you'd put down something heavy — and then your mind immediately tried to pick it back up. The urge to reopen the PR and "just fix one more thing" was almost muscular. You sat with it anyway.
That's the pattern in miniature. The perfectionism isn't really about code quality — you know that now. It's a nervous system response: a learned conviction that the only safe version of your work is the one that can't possibly be criticised. When you short-circuited that loop, your body told you the truth before your mind caught up — it felt like relief.
The Shift: Performance ≠ Safety
We spent the second half of the session connecting this to the bigger picture. The rewrite loop at work and the emotional guardedness with Sarah aren't different problems — they're the same operating system running in different contexts. Both say: if I can be critiqued, I'm not safe. What's changing is that you're starting to test that belief with real data, and the data keeps coming back the same way — the unpolished version is fine. More than fine. It gets the same response, and it costs you dramatically less.
You also made a connection I want you to hold onto: you described your father's approval as something that "weights" your decisions even now — not as a voice in your head, but as a felt sense, a kind of gravitational pull toward doing things the "right" way. That's a significant insight. We're not trying to eliminate that pull. We're building your capacity to feel it and choose differently anyway.
Action Steps:
- Run the experiment again at work. Same protocol — submit after the second review. This time, pay attention to whether the anxiety spike is shorter or less intense than last week. We're building a trend line, not chasing a single data point.
- Try one "unpolished" moment socially. Share an opinion you haven't fully rehearsed — at dinner, in a group chat, wherever feels low-stakes. Notice what your body does before, during, and after.
- Journal the gravitational pull. When you notice your father's "weight" showing up in a decision this week, write down what happened, what the pull was toward, and what you chose. Don't judge it. Just record the data.
The Role of AI in My Practice
This level of documentation — the detailed assessments, the weekly narrative summaries, the progress reports that track real change over time — would be impossible to sustain by hand. I’d either burn out writing them or stop offering them, which is what most therapists do.
AI makes it possible. But the way I use it matters, and I want to be transparent about exactly what’s happening.
What the AI does
Every session is transcribed and processed through secure, healthcare-governed AI services. The AI drafts the session summaries, synthesizes assessment data into the comprehensive report, and compiles the progress reports. It handles the documentation layer — the time-intensive work of turning hours of conversation into clear, structured, useful writing.
What the AI does not do
The AI does not interpret your experience. It does not decide what matters. It does not design your treatment plan or choose what to focus on next. All of the clinical thinking — the formulation, the pattern recognition, the judgment calls about what a moment in session actually meant — that’s mine. The AI is a drafting tool. I’m the clinician.
How your data is protected
Your transcripts are processed through enterprise healthcare agreements I hold with Google and Microsoft — the same class of agreements that hospitals and health systems use. Your data is:
- Encrypted in transit and at rest
- Never stored in any public chat or AI database
- Never used to train any AI model
- Processed under healthcare-grade compliance frameworks (I have detailed documentation available on request)
This is not your words being typed into ChatGPT. It’s a purpose-built clinical workflow running on secured infrastructure.
Why I think this matters
The engagement and outcomes for my clients have improved dramatically since I started working this way. The detailed assessment gives us a shared foundation from day one. The weekly summaries keep the work alive between sessions — clients tell me they reread them, that insights land differently the second time, that having the action steps in writing changes whether they actually follow through. The progress reports catch patterns that are easy to miss in the flow of week-to-week conversation.
Without these tools, we’d be working with significantly less visibility into your progress. I’ve found that makes a real difference in outcomes over time.
What if you’d rather not?
That’s completely fine. We can take a purely conversational approach where I track your progress through questionnaires alone — no transcript processing, no AI involvement. You’d still get my full attention and clinical judgment in session, and the questionnaire data would still give us a useful signal.
I want to be honest about the trade-off, though: you’d be getting less documentation, less between-session reinforcement, and less granular tracking of your progress. For what it’s worth, it also ends up being less work on my end, as I’d have much less material to review between sessions. I’d rather we start on a foundation where you feel fully at ease — but I’d encourage you to consider the full process. The safeguards are robust, and the clinical value is substantial.
The Clinical Thinking Is Mine
I want to be clear about something: the AI does not practise psychology. It processes language. It drafts documents. It is very good at that.
But the moment in session where I notice that your jaw just tightened, or that you changed the subject right when we got close to something — that’s not algorithmic. The decision to sit in silence rather than fill it, to push harder on a particular thread, to reframe a pattern you’ve described a dozen times in a way that finally lands — that’s thirty thousand hours of clinical work, not a language model.
The AI gives me leverage. The craft is mine.
Interested?
If this way of working resonates with you, I’d love to hear from you.