Your Clients Are Already Using AI. Here’s How ThroughLine Is Better.
The question isn't whether AI is showing up between sessions. It's what the AI is designed for.
When we were building ThroughLine, we spoke to a lot of therapists. One thing came up again and again: they already knew their clients were using large language models between sessions. Not as a surprise, but an acknowledgement of concerns they were already sitting with.
Those concerns were consistent: that general-purpose AI is built to answer rather than sit with discomfort, that it can reinforce the desire for an immediate fix rather than building a client's capacity to work through something, and that without relational context it cannot help someone apply what it offers to their own situation.
That makes sense. The gap between sessions is real, and these tools are free and available at any hour. The question is not whether AI is showing up in that space. It already is. The question is what the AI your clients are using was actually designed for.
General-purpose AI is built to be helpful. In most contexts, that is exactly what you want. But in a therapy-adjacent context, "helpful" can mean the wrong thing. A 2025 study from Stanford's Institute for Human-Centered AI compared how general-purpose chatbots and licensed therapists respond to mental health scenarios. The researchers were direct in their conclusion: general-purpose chatbots are unsuitable for safely engaging in mental health conversations, particularly in crisis situations. A tool built to be generally helpful is not the same as a tool built to be safe in this context. A tool optimized to respond thoroughly will advise when it should hold back. It will interpret when it should reflect. It will reassure when reassurance is not what the moment needs. It will fill space that should stay open. None of that is a flaw. It is just not designed around the therapy cycle.
ThroughLine was. Here are three design choices that illustrate what that actually means.
It does not give advice, and it does not interpret.
ThroughLine's core design principle is that it is a mirror, not a therapist. When a user shares something, the app supports their reflection rather than directing it. It does not draw conclusions about what feelings mean. It does not suggest what someone should do or how they should feel. It does not offer reassurance that positions itself as a comfort separate from the therapeutic relationship. The meaning-making belongs to the user and, where it is ready to go, to the session.
This is not a limitation of the technology. It is a deliberate design position. An AI that interprets is no longer a mirror. It becomes an authority. And in a mental health context, that is a role it should not occupy.
It knows when to step back entirely.
ThroughLine has a crisis protocol. If a user shares something that signals they may be at risk, the app does not continue the conversation. It directs the user to real support. There is also a hard design boundary around ideation: ThroughLine does not engage with content involving self-harm or harm to others, regardless of framing. These are not edge cases handled reluctantly. They were designed for from the start, because a tool operating in this space has to know where its role ends.
It has a usage cap, and it is there for a reason.
ThroughLine limits how much a user can engage with the app each day. That cap exists to prevent rumination. Returning repeatedly to difficult emotional content without processing it forward is not helpful, and an AI with no limits on engagement can inadvertently enable that loop. The cap is a design choice that puts clinical considerations ahead of engagement metrics.
Those three things describe something meaningfully different from a general-purpose AI. But this is not just about what ThroughLine does not do. Early users tell us three things consistently.
It helps people feel more active in their own process.
Rather than showing up to sessions hoping the right things surface, users describe feeling more engaged with what is happening for them between appointments. Therapy becomes something they are participating in, not just attending.
It provides structure when sitting down to reflect is hard.
Not everyone finds it easy to start without a prompt. One early user told us the app helped them arrive at sessions knowing where they wanted to start. Having somewhere to go, with a question already waiting, removes the friction that often gets in the way.
It reduces the burden of holding everything in your head.
For people managing busy lives, high stress, and a tendency to overthink, carrying everything until the next appointment is its own kind of weight. One early user described it this way: "My sessions are two weeks apart, and I might not remember how I'm feeling in this moment. I very clearly have a place I can put this, and I know when I go back to it, it's right there, ready and waiting for me."
That is not a feature. That is what it feels like when the gap has somewhere to go.
ThroughLine was built in conversation with therapists, not around them. The design choices above came directly from those conversations, and we continue to develop with practitioner input at the center. If you are seeing clients use AI tools between sessions and want to understand how ThroughLine approaches that space differently, we would love to share more. You can learn about how it works for practitioners at throughlineapp.com, or reach out to us directly at hello@throughlineapp.com.
Reference
Scholich T, Barr M, Wiltsey Stirman S, Raj S. A Comparison of Responses from Human Therapists and Large Language Model–Based Chatbots to Assess Therapeutic Communication: Mixed Methods Study. JMIR Mental Health. 2025;12:e69709. https://doi.org/10.2196/69709
