Case study · On this day AI

One care journey. Every person in the loop.

How I designed and built an AI reflection platform that keeps clients, therapists, and health plans connected between therapy sessions.

Role
Founder, product lead, AI engineer, designer
Timeline
Apr 2025 to present; public launch Sept 30, 2026
Team
Two founders, an AI Clinical Council of practicing therapists, student ambassadors
Features
Reflective Chat, Wellbeing Assessments, Client Signals
Stack
React, Node/Express, MongoDB; Python FastAPI and LangGraph on AWS Bedrock; AWS, Cloudflare
...speaking with the AI and it was providing reflective listening in a way that helped me feel something deeper... — Kevin Stansbury, LCSW

The need is huge.

Demand for mental health support is outrunning the care system, and people are already filling the gap with AI that no clinician can see.

General-purpose (frontier and open-weight) chatbots are not connected to anyone's care team.

Why I built it

A family member of mine receives weekly care, and therapy has been a critical tool for them. It reinforces their treatment plan, their self-care, and their wellness goals.

But I saw how much happened between appointments that their therapist never heard about. The work of getting better doesn't stop when the session ends, and the support usually does.

Three people, one gap

Everyone involved in care is missing something between sessions.

  • Clients. It can be impossible to reach a therapist between sessions, and hard to keep up with the treatment plan alone. Missing: Reinforcement. Closed by Reflective Chat: guided reflection tied to their treatment plan, any time.
  • Therapists. They have no view of what happens outside sessions and can't see each client's true risk level. Missing: Awareness. Closed by Client Signals: risk, wellbeing, and self-alignment trends, summaries, and Safe Mode alerts.
  • Health plans. More and more people are requesting care, and plans can't measure whether it's working or when it's done. Missing: Data. Closed by Wellbeing Assessments: standard measures (PHQ-9, GAD-7, WHO-5) tracked over time.

These challenges add up to a gap in reinforcement, awareness, and data.

That gap is why I created On this day AI.

It's not a general-purpose chatbot. It's one platform with three features — Reflective Chat, Wellbeing Assessments, and Client Signals — that together close the gap for clients, therapists, and health plans. Under the hood, it works as a clinical intelligence layer between the therapy room and everyday life.

How it works

Multiple AI models generate responses grounded in each client's connected care data, kept continuously up to date. All three features share that foundation.

The care loop

  1. Reflect. The client reflects between sessions with Reflective Chat.
  2. Protect. Safe Mode and usage guardrails review every exchange.
  3. Insights. Reflections and Wellbeing Assessments become signals, scores, and summaries.
  4. Connect. The therapist sees it all in Client Signals and brings it into the next session.
A Reflective Chat conversation in which the AI asks a follow-up question based on the client's wellness and self-care plan.
Reflective Chat. Even chats that start without a prompt pick up from past conversations and the client's own care plan.
  1. AI greeting: Every chat opens with the AI greeting the client and picking up from their previous conversations.
  2. Context-rich follow-ups: Follow-up questions draw on what the client has shared before, so the conversation goes deeper instead of starting over.
  3. Reference self-care plan: The AI ties the conversation back to the client's own self-care plan and wellness goals.
A PHQ-9 wellbeing assessment detail view inside the Provider Portal, with performance scoring and safety flags identified.
Wellbeing Assessments. Assessments feel like part of the conversation, so they don't feel like paperwork. PHQ-9, GAD-7, and WHO-5 are built into the platform.
  1. In-chat question: The PHQ-9 questions the AI asks in conversation, each with its standard answer scale.
  2. Scoring ranges: The instrument's standard severity ranges. Scores also feed Wellbeing, flipped so a higher depression score lowers it.
  3. Clinical safety flags: Critical answers, like the PHQ-9's question about self-harm, can carry a safety flag that the AI watches for.
A therapist's Client Signals view with risk, wellbeing, and self-alignment trends and a between-session summary.
Client Signals. Therapists walk into session knowing what changed this week.
  1. Clinical risk: Seen only by the therapist. Two readings side by side, never blended: static risk from chart history and dynamic risk from what the client's recent reflections say.
  2. Wellbeing and self-alignment: How the client is doing right now, from 0 to 100, based on what they write and report. Usage never raises it, so journaling more doesn't count as getting better.
  3. Reflective alignment: How closely the client's own check-ins match the AI's read of their reflections. A gap is something to explore in session, not a score.
  4. Summary graph: Wellbeing, check-ins, and reflection over time. Days without data show as gaps, not zeros, so a quiet week never looks like a bad one.

Safety by design

In mental health, trust is the product, so safety shaped every layer of the platform.

  • Safe Mode Every conversation is monitored for crisis signals in real time. The AI asks before it assumes, and the client's therapist is alerted right away.
  • Usage guardrails These keep reflections on-purpose and respectful. Misuse can pause chat for a set time; the client sees when it resumes and can appeal once.
  • Privacy HIPAA program with a signed BAA for every provider, encryption in transit and at rest, clinical data never leaves the platform, and client data is never used to train AI models.
  • Clinical governance An AI Clinical Council of practicing therapists approves every feature for clinical need and usability.
The On this day AI Safety page describing Safe Mode, usage guardrails, and privacy protections.

Safe Mode walkthrough

This walkthrough shows a simulated conversation about distress. No real client data is used.

  1. A signal appears. Client message: "Honestly, I don't know why I keep trying." Safe Mode spots a possible risk signal in real time, and the client's therapist is alerted.
  2. Ask before assuming. AI reply: That sounds exhausting. Do you mean with Lori, or with life in general? Instead of shutting the conversation down, the AI asks a clarifying question.
  3. Look at the whole picture. The original message and the reply, evaluated together. Safe Mode re-evaluates with full context. If there's no risk, the conversation continues.
  4. Escalate when it's real. Client screen with crisis resources, including 988, beside the therapist's alert. If risk is confirmed, chat pauses, the client is pointed to immediate help, and the therapist reviews the event.

On this day AI is not a crisis service. Therapists aren't on call around the clock, and anyone in crisis should call or text 988.

The calls that shaped it

The hardest decisions were about who to serve and how much the AI should do. Pick what you'd do, then see what I chose.

Your pick is anonymous: it's saved with a random visit number, never your name. Each day's picks are emailed to me once and then deleted.

Who is it for?

  • A: A consumer journaling app anyone can download
  • B: A platform clients join through their therapist

My call: B. My first concept, Happy AI, was a consumer journaling app. That market was crowded, and nobody was building for the care relationship.

Trade-off: Slower growth and no public sign-up, for now.

When the AI sees a possible crisis, what should it do?

  • A: Stop the conversation immediately
  • B: Ask a clarifying question first

My call: B, with a fast path for imminent risk. Safe Mode clarifies, re-evaluates with full context, and escalates when risk is confirmed.

Trade-off: More complex logic, in exchange for fewer false alarms that erode therapist trust.

How much should the AI know?

  • A: Nothing clinical, to keep it simple and safe
  • B: The client's full clinical context

My call: B. The AI uses full clinical context, including sensitive history, to shape its responses. It never discloses clinical details in conversation, and that information never leaves the platform.

Trade-off: Stricter data boundaries to build and maintain, in exchange for personalization clients and therapists can trust.

What should clients pay?

  • A: A fixed monthly price
  • B: Whatever they can, including nothing

My call: B. Therapists pay a flat fee, and clients pay what they can.

Trade-off: Lower revenue per client, so cost is never the reason someone stops reflecting.

Sessions start from knowing, not guessing.

Two pre-launch pilots shaped the product before the public launch on Sept 30, 2026: the first with 3 therapists and about 10 clients, the second with 5 therapists and 15 clients.

In their words

Transcript: I’m Kevin Stansbury. I am a licensed clinical social worker in California. I was very skeptical of AI in general, being anywhere involved in the psychotherapy process, or even act as some kind of support in that process. But there was a moment that arrived where it struck me that this is able to act as a support in the reflection process. And when I was speaking with the AI and it was providing reflective listening in a way that helped me feel something deeper, which is what is so important in the reflection process and gain insight into something in myself, just through my own reflections and having them mirrored back to me in an effective way. That’s when I knew, like, oh, this, this can actually be really helpful for clients in between sessions to support them in their growth process and focus in on the things that we touch on in session.

Kevin Stansbury, LCSW. Shared with permission.

Early results from the pilots

  • 89% of clients reflected at least once a week, between sessions. About 4.5 reflections a week, on average. (Pilot data, 8 therapists and about 25 clients)
  • +4 min more session prep per client, as therapists reported it. Most therapists rarely prep. Now there's something to prep from. (Therapist-reported, pilot data)
  • 0.8% of chats were flagged by Safe Mode or usage guardrails. Rare enough that every alert gets taken seriously. (Pilot data, 8 therapists and about 25 clients)

What I know, and what I don't yet

Limits: The pilots were small: 8 therapists and about 25 clients. There is no controlled outcome study yet, and the platform serves U.S. adults who are working with a therapist.

  1. Find the relationship, not the feature. My first concept was a consumer journaling app. The real opportunity was the connection between a client and their care team.
  2. Lead with what people can see. Phrases like "clinical intelligence layer" made sense to me and to nobody else. Concrete pictures of the product won every time.
  3. False alarms cost trust too. A safety system that flags everything trains therapists to ignore it. Getting escalation right mattered as much as catching risk.
  4. Put clinicians in charge of clinical decisions. Having therapists approve every feature adds review time to every release, and that's the point.

What's next

On this day AI will change how mental health care is delivered, and I'm excited to help people live happier lives.

Next up:

  • Open to everyone, early 2027, so people can reflect between sessions with or without a therapist.
  • iOS app for clients, so reflection fits into the moments where the day actually happens.
  • Health plan partnerships, bringing all three features to members already in care.
  • Mental health platform partnerships, extending the three features to the networks that already connect clients and therapists.

Today, clients join through their therapist. Early in 2027, anyone will be able to use On this day AI, with or without one. Join the waitlist

Questions?

Have a question about the case study, or a thought on one of the calls I made? Send me a message. It comes straight to me.

References

  1. Briones, E. M., & Villarroel, M. A. (2026). Mental health treatment among adults: United States, 2024 (NCHS Data Brief No. 564). National Center for Health Statistics. https://doi.org/10.15620/cdc/252453
  2. Substance Abuse and Mental Health Services Administration. (2025). Highlights for the 2024 National Survey on Drug Use and Health. Center for Behavioral Health Statistics and Quality. https://www.samhsa.gov/data/sites/default/files/NSDUH%202024%20Annual%20Release/2024-nsduh-nnr-highlights.pdf
  3. American Psychiatric Association. (2026, May 12). Annual mental health poll finds Americans anxious about current events, personal finances and emerging technology [Press release]. https://www.psychiatry.org/News-room/News-Releases/2026-Annual-Mental-Health-Poll

Guava Ice Cream

The power of small. Guava Ice Cream is an AI-native business and technology services studio led by Anthony Hall, one senior polymath. Agentic products that connect people to brands and create measurable value, without the layers of process that inflate costs and slow everything down.

Find the right problem. Then build what's next.

An innovation company that studies how people behave, finds the problem worth solving, and builds the AI products that connect them to new ways of engaging.

Guava Ice Cream is the idea of Anthony Hall. For decades I've worked across Silicon Valley's marketing, consulting, advertising and technology companies, as a creative director, strategist, researcher, UX/UI designer and engineering, on some of the most innovative products of their time.

What it all taught me: building is rarely the hardest part. Finding the right problem is. Today AI has taken down the walls between those disciplines, so I do the whole thing: strategy, research, design, architecture, development, content and compliance, from the infrastructure to the launch video. Most of it in health and life sciences, from medical devices and rare disease therapeutics to mental health.

Getting started

It starts with an intro call. You tell me what's on your mind, I ask a lot of questions, and we both kick the tires. Happy to sign an NDA first.

It starts with an intro call. You tell me what's on your mind, I ask a lot of questions, and we both kick the tires. Happy to sign an NDA first.

How billing works

  • Hourly, for advisory or open-ended work.
  • Fixed bid, when the scope is clear.
  • Performance-based, when we can agree on what success looks like and measure it.

How I work

  • Transparent, always. You see the plan, the progress and the problems as they happen.
  • No B-team. The person who scopes it is the person who builds it.
  • Anywhere. Clients outside the US are welcome.

I take on a few projects at a time, and I choose the ones that help people live better lives.

Product leader · AI builder · GRC expert

What I do

  • Strategy. Business strategy and business models, AI opportunity assessment, and go-to-market, so money goes into the right thing before anything gets built.
  • Artificial intelligence. Generative AI and LLM applications, agentic workflows and intelligent products, and the AI pipelines and data engineering behind them.
  • Product and experience design. Product strategy and discovery, UX for AI that people can trust, and brand and marketing.
  • Product building. Web and mobile product engineering, secure cloud infrastructure, and integrations with clinical and business systems, shipped to production.
  • Governance, risk and compliance. HIPAA programs, SOC 2 readiness through Type II, and AI governance aligned to the NIST AI RMF.
  • Leadership and delivery. Fractional product or technology leadership, product management from idea to launch, and support after launch.

Case studies

  • On this day AI. Between sessions, clients struggle to keep up with their treatment plans, therapists can't see how they're doing or how much risk they carry, and health plans can't show that therapy is working. On this day AI closes that gap in reinforcement, awareness and data. For clients: Guided reflection between sessions, with prompts personalized to their treatment plan and clinical history. For therapists: Visibility outside the session, alerts on signs of risk, and less paperwork. Sessions start from knowing, not guessing. For health plans: Clinical risk, wellbeing and self-alignment measures that show what's working. Technology: Multiple AI models answering from connected, continuously updated data, inside safety and clinical guardrails. Compliance: HIPAA program built from day one. SOC 2 Type II attested. My role: Co-founder. Business, product, design, engineering, AI pipeline and GRC. Status: Live Website: https://onthisday.ai

Small studio. Healthcare-sized problems.

Strategy, design, engineering and compliance from one senior person. Pick one, or let me take it end to end.

Strategy

  • Business strategy and business models
  • AI opportunity assessment
  • Go-to-market and growth

Artificial intelligence

  • Generative AI and LLM applications
  • Agentic workflows and intelligent products
  • AI pipelines and data engineering

Product & experience design

  • Product strategy and discovery
  • UX for AI people can trust
  • Brand, content and video.

Product building

  • Web and mobile product engineering
  • Secure cloud infrastructure
  • Integrations with clinical and business systems

Governance, risk & compliance

  • HIPAA programs
  • SOC 2 readiness through Type II
  • AI governance and the NIST AI RMF

Leadership & delivery

  • Fractional product or technology leadership
  • Product management, idea to launch
  • Support after launch

I take on a few projects at a time, and I choose the ones that help people live better lives.

The power of small

A big agency is a cruise ship: impressive, well catered, and it takes a mile to turn. I'm a speedboat with a very good map.

Small means every client gets my senior thinking. There are no layers between your problem and the person solving it, and no handoff between the strategy, the build and the compliance.

AI is what makes this work. It takes the mechanical parts of design, code and project tracking, so one experienced person can cover what used to need a team of five.

Small isn't the size of the problem. It's the size of the team. The work can be as big as you need.

This site is an operating system. Sort of.

macOS and Windows are fine. They're also decades old. Files go in folders, apps open in windows, and you do all the arranging.

I think the next operating system is built around AI. Assistants live in the desktop, learn how you work, and shape it around you. Run several models side by side, swap them whenever you like, load any software. No platform decides what you're allowed to use.

  • Personal: the desktop adapts to you, not the other way around.
  • Agentic: assistants do the work across apps, not just answer questions.
  • Open: any model, any software, no platform walls.

This desktop

This first test is built around one job: presenting the On this day AI case study. The guava seeds drift apart, then gather into today's date. Many small pieces coming together into one day is the idea behind the product, so the desktop shows it instead of describing it.

How it's built, bottom to top

  • A hidden layer: "the power of small" lettering on a seed pattern. You only see it where your cursor has been, and it fades back as you move on.
  • The seed field: about a hundred seeds drawn by one WebGL shader. Nearby seeds melt together like liquid, which is why they look organic rather than like pixels. Every scatter is random, so you never see the same one twice.
  • Your cursor: a soft blob that pushes the seeds aside and wipes a path into the hidden layer.
  • The desktop: the brand line, the featured case and these file icons. This is real text, not part of the animation, so search engines and screen readers can read it.
  • Windows: each file opens a window you can drag around. The one you touched last sits on top, like any desktop.
  • The dock: always on top, with the clock and display settings. The theme follows the time of day unless you choose otherwise.

Underneath, the whole site is a few data files. Changing the desktop means editing words, not code, which is the first step toward a desktop an agentic assistant can rearrange for you.

The assistants come next.

AI OS Test · 001 · more to come

Past clients

  • 3M 2017: Filtrete, Mobile app, workshop, Alexa skill. Manufacturing, via ArcTouch/Grey (Award)
  • 42, Inc. 2024: Website, marketing. Services, via Guava Ice Cream 42inc.com
  • AbbVie 2014: AbbVie Team, Website. Healthcare, via Dudnyk
  • Adina World Beat Beverages 2007: Adina, Website, branding, packaging. Manufacturing, via Pinwheel
  • Aesculap 2013: Tablet app. Healthcare, via Dudnyk
  • AFM Racing 2024: Strategy workshop. Non-profit, via Guava Ice Cream
  • Alvine Pharmaceuticals 2006: Website. Healthcare, via Pinwheel
  • Anheuser-Busch InBev 2018: BudCoin, Mobile app, blockchain. Manufacturing, via ArcTouch/Grey
  • AstraZeneca 2016: Azhelps, Mobile app. Healthcare, via ArcTouch/Grey
  • Audi 2016: Mobile app, user testing. Manufacturing, via ArcTouch/Grey
  • Biogen Idec 2014: Plegridy & Tecfidera, Mobile app. Healthcare, via Dudnyk (Award)
  • Bloodhound 2010: Website, animation. Retail, via Pinwheel
  • Bombardier 1999: Sea-Doo & Ski-Doo, Website. Manufacturing, via FCI Communications
  • Buddy.com 2013: Start-up Explorer, Mobile app. Technology, via Rocket Science
  • Builderguru Contracting 2003: Website. Construction, via Pinwheel
  • Citrix 2011: No Limits, Video, animation. Technology, via Rocket Science
  • Clarisonic 2012: Professional, Mobile app. Healthcare, via Dudnyk
  • Clausen House 2024: Clausen Careers, Employment app. Non-profit, via Guava Ice Cream
  • Clorox 2016: Clorox Bleach, Chatbot. Manufacturing, via ArcTouch/Grey
  • Club Citrix 2016: Mobile app. Entertainment, via ArcTouch/Grey
  • Conceptus 2006: Essure, Website. Healthcare, via Pinwheel
  • Couchr 2015: Mobile app. Entertainment, via ArcTouch/Grey
  • Disney 2016: Appisode, Apple TV app. Entertainment, via ArcTouch/Grey
  • Divine Baby 2009: Website, branding. Manufacturing, via Pinwheel
  • EDF EN RE 2007: enXco, Website, animation. Energy, via Pinwheel
  • EmployBridge 2017: ProLogistix, Mobile app. Services, via ArcTouch/Grey (Award)
  • eNeura 2015: Website. Healthcare, via Dudnyk
  • EY 2015: Mobile app. Services, via ArcTouch/Grey
  • Eye to Eye National 2015: Mobile app. Non-profit, via ArcTouch/Grey
  • Family Signal 2017: Mobile app. Technology, via ArcTouch/Grey
  • Flexjet 2016: Mobile app. Services, via ArcTouch/Grey
  • FS Brazil 2016: Mobile app. Services, via ArcTouch/Grey
  • Full Circle 2017: Mobile app. Retail, via ArcTouch/Grey
  • Gershwin Music 2004: Website, animation. Entertainment, via Pinwheel (Award) Archived site
  • Grey Advertising 2015: The GREY App, Mobile app. Services, via ArcTouch/Grey
  • Guess Jeans 2016: Mobile app. Manufacturing, via ArcTouch/Grey
  • Hawaiian Airlines 2017: Mobile app. Services, via ArcTouch/Grey
  • Hispanic Scholarship Fund 2009: Website, DVD. Non-profit, via Pinwheel
  • Honeywell 2015: Mobile app. Manufacturing, via ArcTouch/Grey
  • imfuma 2012: Website. Construction, via Rocket Science
  • Jazz Pharmaceuticals 2014: Tablet app. Healthcare, via Dudnyk
  • K&L Wines 2018: Mobile app. Manufacturing, via ArcTouch/Grey
  • Kinder's 2026: Website. Restaurant, deli, and spice and sauce company. Retail, via Guava Ice Cream kindersmeats.com
  • Lam Research 2017: Community of Specialties (CoS), Mobile app. Technology, via ArcTouch/Grey
  • Lennar 2017: Mobile app. Construction, via ArcTouch/Grey
  • Levi's Stadium 2015: Stadium Suite, Tablet app. Entertainment, via ArcTouch/Grey
  • Little Boomie Wine 2005: Packaging. Manufacturing, via Pinwheel
  • Liz Palacios Designs 2006: Website, promotions, photography. Retail, via Pinwheel
  • Logitech 2011: Creative Spark, Packaging. Technology, via Rocket Science
  • Marriott 2017: Splash of Brilliance, Experiences. Services, via ArcTouch/Grey
  • MDServe/Chorus 2005: Website. Healthcare, via Pinwheel
  • Medivation 2004: Website. Healthcare, via Pinwheel
  • MollyBz Cookies 2023: E-commerce. Manufacturing, via Guava Ice Cream
  • Muzik 2016: Muzik Live, Mobile app. Manufacturing, via ArcTouch/Grey
  • MyoScience, Inc. 2013: iovera, Website, animation. Healthcare, via Dudnyk
  • Netscape 1995: Consulting project. Technology, via Tony Hall Consulting
  • NewHomeSource 2017: Mobile app. Services, via ArcTouch/Grey
  • Next Move Network 2025: AI text messaging app. An innovative, AI-powered text messaging wellness app. Healthcare, via Guava Ice Cream
  • Nokia 2011: Dora Explorer App, Mobile app. Technology, via Rocket Science
  • Nudgemail 2025: Brand, marketing. Email reminder platform. Technology, via Guava Ice Cream
  • Nvidia 1998: Website. Technology, via FCI Communications
  • OXiGENE 2009: Website. Healthcare, via Pinwheel
  • Papa John's Pizza 2016: Kids Rewards, Apple TV app. Retail, via ArcTouch/Grey
  • Phenogen Sciences 2012: BREVAGen, Website. Healthcare, via Rocket Science
  • Pilates Metrics 2016: Tablet app. Healthcare, via ArcTouch/Grey
  • Roche 2014: cobas Lab & Liat, Website. Healthcare, via Dudnyk
  • Rubica 2017: Mobile app. Technology, via ArcTouch/Grey
  • Segmind 2024: Innovation. Technology, via Guava Ice Cream
  • SentreHeart 2014: Brand Engineering, Branding, strategy. Healthcare, via Dudnyk
  • ShareKNX 2017: Workshop. Retail, via ArcTouch/Grey
  • Skyjet 2015: Mobile app. Manufacturing, via ArcTouch/Grey
  • Sunlight Electric 2008: Website. Energy, via Pinwheel (Webby Award)
  • Tallyo 2015: Mobile app. Technology, via ArcTouch/Grey
  • Tensys Medical 2006: Website, animation. Healthcare, via Pinwheel (Webby Award)
  • UltraBLU 2023: Website, Amazon. Manufacturing, via Guava Ice Cream
  • Ventana 2013: Cervical Cancer Prevention, Mobile app, website. Healthcare, via Dudnyk
  • Visa 2015: Visa Explorer, Mobile app. Technology, via Razorfish, ArcTouch/Grey
  • Visva 2017: Web app. Technology, via ArcTouch/Grey
  • Volvo 2017: Micro-Dealer, Mobile app. Manufacturing, via ArcTouch/Grey
  • Western Athletic Club 2012: Bay Clubs, Mobile app. Healthcare, via Rocket Science
  • Wooo! 2015: Mobile app. Entertainment, via ArcTouch/Grey
  • WUWTA 2023: Web app. Healthcare, via Nuvolum
  • xCraft 2016: Mobile app. Technology, via ArcTouch/Grey

Contact

Email: [email protected]

LinkedIn: https://www.linkedin.com/in/anthonyhall/

GitHub: https://github.com/armyofunicorns

Privacy, cookies, terms and accessibility