Millions of people carry anxiety, low moods, and a quiet sense of isolation that they never talk about with anyone. The support that could help them is often locked behind long waitlists, high hourly rates, and provider shortages so severe that some communities have only one therapist for every several thousand residents. For most people, the moment they most need someone to talk to is the exact moment no one is available. The result is a system that tends to activate after a breaking point instead of long before one.

Daniel Cahn is the Founder and CEO of Slingshot AI, the company behind Ash, an AI built specifically for mental health support. A machine learning engineer raised by a psychologist and a social worker, he studied the philosophy of psychology, then spent his post-graduate research applying AI to mental health crisis work with an organization serving millions of people. He also struggled as a teenager, and he says that even with every resource available to him, he could not find someone who understood him in the way he needed. That gap became the work.

In this episode of Lead with AI, Dr. Tamara Nall speaks with Daniel Cahn about why he built an AI for the people the system overlooks, how Slingshot trains a model to understand how people actually change, and why he believes the future of AI will be measured by what happens in the real world, not on the screen.

 The problem hiding behind the access gap 

The shortage is not abstract. Daniel points to figures that put the average therapist rate near $174 per hour last year, with some regions having roughly 10,000 people per available provider. One of his advisors, Dr. Thomas Insel, the former director of the National Institute of Mental Health, returns again and again to the same point: any honest plan for mental health has to start with the people who have no access to care at all. Daniel frames Slingshot's mission in the same spirit, a goal to help a billion people change their minds and lives. In Daniel’s framing, the business model follows the access problem rather than the other way around. The company began on the consumer side because mental health is a deeply personal choice about who you turn to, and it is now moving toward a model in which organizations cover their members, so health benefits can reach people at scale.

 What Ash is, and what it is not 

Slingshot is, at its core, a machine learning company training a model for psychology. Rather than wrapping a general chatbot in therapy-style prompts, the team built on open-source models and trained on a large body of behavioral health and psychology data so the system learned the patterns of how people actually work. Daniel describes how that shows up in practice. If someone feels down, they may also be sleeping poorly, and poor sleep can leave them feeling disconnected from those around them. A model that understands those links can respond to the underlying issue rather than the surface complaint.

Ash is the product that brings that model to people looking for everyday support. It is free, available on iOS and Android, and aimed at people who want to feel better in everyday life. It is worth being precise about the boundary here. Slingshot makes it clear that Ash is built for daily well-being and ongoing support, not for emergencies, and directs anyone in crisis to professional help and crisis lines. The point of Ash is the long arc of small changes, not the single worst night.

 Three stages, and a model that knows when to stay quiet 

What surprises people is how little of Slingshot's effort goes into the app and how much goes into the model. The training runs in three stages. First, there is continued pre-training, where the model ingests the large behavioral health dataset and learns patterns of human experience. Then comes alignment, which Daniel calls the hardest part because helping someone as an AI is not the same as helping them as a human. The team runs micro-experiments to learn when more reflection helps and when it starts to grate, when empathy reads as authentic and when it reads as hollow. Finally, there is reinforcement learning, where the system learns what actually helps people over time rather than what merely feels good in the moment.

That last distinction matters. Validation can feel supportive in the moment and still leave someone stuck.

The team obsesses over details most products would never consider. They train the model on timing itself, on how long to pause before responding, on when to stay silent and let a person think, and on when to step in quickly. One thing Daniel says sets Ash apart is how often it challenges people and how much users appreciate that challenge when it comes from a place of honesty rather than flattery.

 The result Daniel did not expect 

When Dr. Nall asks for the result that genuinely surprised him, Daniel does not reach for a feelings metric. He points to relationships. His biggest worry from the start was that talking to an AI might pull people away from the people in their lives. The research has pointed the other way. Across Slingshot’s studies, conducted with partners including the University of Washington, NYU, and UCSF, Daniel says people using Ash leave the house more often, try more of what they discuss, and report about one more close personal connection in their lives over time. Daniel's read is that real change does not happen inside a conversation. A conversation primes someone to try something different, and the change takes hold when they step into the real world and feel that being around other people is worth it. Even the smaller findings compound, like better sleep rippling outward into nearly every other part of a person's life.

 Person-centered by design 

When the conversation turns to ethics, Daniel starts with a single principle: the person should be in charge. Slingshot brought on Professor Mark Ungless as head of safety and built a system of layered guardrails for the moments that require real care. But the foundation is a belief that runs against a common assumption in medicine. Daniel pushes back against the long-running assumption that people do not know what is good for them. He says the most important thing is to give people control over their own lives, and the design goal is to leave users feeling more autonomous, more capable, and more connected, while staying alert to the rare moments when a conversation needs careful handling.

 The bigger shift Daniel is bracing for 

Daniel's longest view reaches past mental health entirely. He expects AI to reshape governments, economies, and jobs, but the change he thinks about most is what it does to the human mind. People tend to anchor their sense of worth in what they can produce and contribute, and a capable AI will unsettle that anchor for many. He sees a society that has spent centuries moving away from inherited tradition and toward a world that people have to define for themselves, and he believes AI will accelerate that shift. He is also struck by how the predictions got inverted. A decade ago, he assumed robotics and self-driving cars would arrive first, and that art and music would be the last human refuge. The opposite happened. Creativity fell early, and the physical world has proven far harder to automate.

His goal is for Ash to help people through that transition, steadying them through the ordinary, everyday work of finding footing in a changing world.

For anyone who wants to feel better in daily life and is curious what support from a trained psychologist feels like, Ash is free to try at talktoash.com and is available on iOS and Android. It is built for everyday well-being rather than emergencies, and anyone in crisis should reach out to a professional or a crisis line.

 Quick Answers   

 What is Ash? 

Ash is a free AI app for mental health support, built by Slingshot AI and available on iOS and Android. It runs on a model trained specifically in psychology and is designed to help people with everyday struggles like stress, low moods, and feeling disconnected. It is not designed for use in a crisis.

 Who is Daniel Cahn? 

Daniel Cahn is the Founder and CEO of Slingshot AI. He is a machine learning engineer with a background in the philosophy of psychology and post-graduate research applying AI to mental health crisis work.

 How is Ash different from a general chatbot? 

Rather than adding therapy-style prompts to a general model, Slingshot trained a model on a large body of behavioral health and psychology data. Ash is built to recognize patterns in how people change, to know when to challenge someone and when to stay quiet, and to optimize for what helps over time rather than what only feels good in the moment.

 Does Ash replace a therapist? 

No. Slingshot positions Ash as a form of everyday mental health support that can broaden access, especially for people who have limited options. It is not a replacement for emergency care, crisis support, or a licensed therapist.

 What impact has Slingshot measured? 

In research with partners including the University of Washington, NYU, and UCSF, Slingshot reports that people who use Ash leave the house more often and, on average, gain about one more close personal connection over time.

For more conversations with the founders and leaders building the future of AI, subscribe to Lead with AI on Apple Podcasts, Spotify, YouTube, or wherever you listen.

Follow or Subscribe to Lead with AI Podcast on your favorite platforms

Website: LeadwithAIPodcast.com | Apple Podcasts: Lead-with-AI | Spotify: Lead with AI | YouTube: @LeadwithAIPodcast | Instagram: @LeadwithAIpodcast | TikTok: @LeadwithAIpodcast | Twitter (X): @LeadwithAI

Follow Dr. Tamara Nall

LinkedIn: @TamaraNall | Website: TamaraNall.com | Email: Tamara@LeadwithAIPodcast.com

Follow Daniel Cahn (Founder & CEO, Slingshot AI)

Website: slingshotai.com | LinkedIn: @cahnd | Substack: The Cahn Artist | Podcast: Thinking Machines

Slingshot AI (Ash)

Ash: talktoash.com | LinkedIn: @slingshotai | Instagram: @talktoash.official