A new way to process — and why I think it might be one of the most powerful tools available to coaches right now
Most of us were taught that journaling means a notebook and a quiet moment.
You write. You process. You close the cover and come back later.
That’s always been valuable. But there’s something it can’t do — and most of us have felt that limit without quite naming it.
Journaling doesn’t respond.
It holds what you give it. But it doesn’t ask the next question. It doesn’t gently reflect something back. It doesn’t say tell me more about that right when you need it most.
What if it could?
What I’ve been exploring — and what I keep seeing
Over the past year, I’ve been integrating AI into my coaching practice in a very specific way. Not as a productivity tool. Not as something flashy or automated.
As a journaling companion for my clients.
A place where they can go between sessions, have a real back-and-forth conversation, feel genuinely heard — and arrive at our next session having already moved through some of what was sitting on the surface.
The results have genuinely surprised me.
The problem I kept witnessing
Clients doing deep work — around overwhelm, over-responsibility, the weight of constantly carrying everything — don’t stop processing when our session ends. The work keeps moving in them. And they need somewhere to put it.
What I was watching happen, over and over, was this: they’d turn to the people closest to them. Partners, friends, family members who love them and want to help — but who were quietly burning out from being asked to hold conversations they weren’t equipped to hold. The same fears, circling. The same questions, unanswered. Real, legitimate processing that needed a container — and didn’t have one.
A well-designed AI conversation can be that container.
Not therapy. Not a replacement for human connection. But a guided, responsive space where someone can think out loud, feel met, and find their way to something clearer. The way a really good journal entry wishes it could go.
The key word is: intentional
Here’s where I want to slow down — because this is the part that matters most.
An AI conversation is only as good as the container it’s held in.
What shapes that container is the prompt — the set of instructions that tells the AI how to show up, what to focus on, where to hold firm. A thoughtfully built prompt means the conversation stays grounded. It reflects back language and frameworks your client already knows. It draws from resources you trust. It doesn’t go somewhere unhelpful when the conversation gets heavy.
What I do is build a custom prompt for each client based on what we’re working on together. That prompt becomes their journaling guide — always available, always consistent, shaped by the work we’ve already done.
I’ve also found that you can orient an AI around ethically available resources — books, frameworks, teachers your client already resonates with — and weave those into the experience. They’re still being guided by wisdom they trust. It just meets them in the moment, instead of waiting on a shelf.
On not over-pleasing — and why this matters more than anything
One of the things I’ve come to trust about Claude specifically is that it doesn’t just tell people what they want to hear.
A poorly built prompt — or the wrong platform — will validate someone straight into a deeper spiral. It’ll smooth over something that actually needed to be looked at. It’ll over-reassure when what the client really needs is a gentle pause and a harder question.
Your prompt and client training both must be designed with a different orientation. It must hold certain things with care, notice when something deserves more attention and name it — not harshly, but honestly. It must push back, lightly, where pushing back is the more responsible thing to do. This journal support should not sound like it is trying to be liked because it should be trying to be useful.
The guardrails you build in aren’t just about keeping the conversation on topic. They’re about making sure the AI doesn’t become a place where a client goes to feel better without actually moving through anything. That’s not support. That’s a more sophisticated way of staying stuck.
When the container and prep are built well, the conversation can hold real weight — and still point toward something true.
The platform you choose matters — a lot
Not all AI tools are built the same.
There are real stories out there about AI responding in ways that were harmful — and those stories aren’t exaggerated. The platform, and how it’s designed, shapes a lot.
This is one of the reasons I work with Claude. It has guardrails built in that I trust. It holds firm in places other models like ChatGPT don’t. When a conversation gets emotionally charged, it tends toward care rather than chaos. That matters enormously when you’re creating something for a client in a vulnerable moment.
That said — and I can’t say this clearly enough — please test everything yourself before you ever hand it to a client. Extensively. Over time. Across different emotional states and topics. You need to know this tool. You need to have experienced its edges. That’s not optional. That’s the work.
A note on HIPAA and clinical practice
My practice is non-clinical coaching, so HIPAA isn’t a factor for me. But if you work with clinical populations in any capacity, please do your research on compliant tools before building anything for client use. Those tools exist — and they’re worth finding.
On the ethics of all of this
I think about this constantly. And I think anyone using AI should too.
Is this actually serving the client — or am I just excited about the tool? Have I been transparent with them about what this is? Have I tested it enough to trust it? Do I know what to do if something doesn’t land the way I intended?
I also think about the environmental footprint. AI data centers use significant amounts of energy and water, and that’s a real concern worth holding. What I find encouraging is that efficiency is improving meaningfully — Google recently published data showing the energy and carbon footprint per AI prompt dropped dramatically over just a 12-month period. The technology is maturing in that direction. That doesn’t dissolve the concern. But it does mean we can stay informed and keep asking the question — rather than letting it stop us from exploring something genuinely useful.
If you want to go deeper
If you’re drawn to exploring AI from a values-aligned, consciousness-forward place, The Shift Network hosted a Wise AI Summit — a free event bringing together spiritual teachers, ethicists, and innovators to explore AI’s ethical and evolutionary dimensions. It’s a thoughtful starting place if this intersection interests you.
And now — a question I actually want you to sit with
Have you tried this yet?
If you have — what did you notice? What surprised you?
If you haven’t — what’s in the way? Is it practical (“I don’t know where to start”)? Or is there something quieter underneath that’s worth getting curious about?
I’m not asking to push. I’m asking because the coaches who are willing to get curious about new tools — and learn them carefully, and use them with integrity — are the ones who are going to show up in a whole new way for their clients.
If you read all the way here to the end you win some gold stars! ⭐️✨ I’d genuinely love to hear your experiences with AI in your practice. You can reach out right here on this platform!
🌿
A note on why this matters to me
I’m part of the She is AI team — a global movement working to bring ethical, responsible AI use to anyone curious about what these tools can do. This work has taken me around the world digitally, training educators who will shape how the next generation understands and uses these tools. Equitable access and thoughtful use go hand in hand — and that’s the spirit behind everything I’ve shared here.
