The same purpose.
A different path
to the product.
I became an investor in MY INNER ME in 2021. In 2026, we closed the business. The app offered preventive self-help for children and families. Here is how I would approach it if I started in 2026 — with AI and the lessons I learned.
A personal reflection on a closed business and a possible approach today. Proposals are hypotheses, not completed AI projects, demonstrated health outcomes or an announced relaunch. This case will be expanded over time.
A meaningful purpose.
A business that closed.
MY INNER ME aimed to make psychological self-help easy and affordable for children and families, with confidence in the quality of its content. The app included audio exercises, a diary and mood check-ins. The child had the need, while the parent played an important role in understanding and supporting them.
The app and its first programmes already existed when I joined. I led further product development, including the move to Flutter as a shared codebase for iOS and Android. I helped expand the topics, establish collaborations with external psychologists, reduce content production costs and establish partnerships with Dansk Sundhedssikring and Danica Pension.
The company has since closed. In my assessment, we focused too much on the product and too little on the commercial side. For too long, I saw my responsibility as primarily product and became involved in sales and negotiations too late. Partnerships brought a more stable customer base, but we started too late and lacked strong data to demonstrate value.
Another important cost was the time we spent standing still because we thought the next step required an expensive new feature. We also believed we needed to produce too much of the content ourselves, without allocating enough psychologist time.
I would start
with anger management.
The need for accessible support remained. The challenge was finding a solution and a sustainable business. Starting today, I would focus on anger management for children and families: our strongest topic based on usage and feedback.
That gives me a concrete starting point, but does not demonstrate a health outcome. I would combine quantitative measurement with qualitative research into whether families understand and can use the offer. I would avoid in-depth interviews about a child’s personal difficulties and draw on professionals’ knowledge of needs and appropriate boundaries.
Can a focused anger-management offer create enough value for a family to pay and for a psychologist to contribute?
First an offer.
Then a simple web app.
I would find a psychologist willing to test the idea with me and build a landing page myself using AI. The offer would explain who it is for, what the family gets and the monthly membership price. Sign-ups would be followed by a test of actual willingness to pay.
I would then start with a mobile-friendly web app. App Store and Google Play distribution would not be prerequisites for learning. The first product would provide a simple route to professionally reviewed content and make the parent’s role clear.
I would explore what psychologists actually need before building a platform for them. Attracting more providers of quality-assured content is a central hypothesis, not something I would take for granted.
The first paying families and professional collaborators would help test the model. A larger membership base is a later goal, not a requirement for starting to learn.
A shorter path
from idea to testing.
I already build websites and processes using AI. In a new MY INNER ME, I would use that experience to build and change the landing page and prototype myself and examine the business model’s assumptions. Specific tools and workflows would be chosen according to the task and the controls required.
| Task | Where AI could help | Responsibility and review |
|---|---|---|
| Test the offer | Draft landing pages, messages and prototypes that can be changed quickly. | I prioritise hypotheses and assess evidence from real users. |
| Build the platform | Draft code and suggest tests for a simple web app. | Technical review of code, login, access and operations before using real family data. |
| Challenge the business model | Structure scenarios for pricing, costs and payments to psychologists. | Assumptions must be checked and tested with families and professionals. |
| Prepare professional material | Transcription, draft translations and organisation of professionals’ material. | Professionals review and approve the version published. |
AI should free up time and reduce the cost of exploring a direction. That is a hypothesis about a better workflow, not a demonstrated promise of savings or more families helped. I would still seek technical and data-protection expertise for areas I cannot take responsibility for myself.
People remain responsible
for professional quality.
Content would come from psychologists and relevant professionals. I would explore a model in which one or more psychologists review each other’s work before publication. AI could help with preparation, but would not replace professional judgement.
A chatbot or open-ended AI guidance would not be part of the first version. The initial task is to make quality-assured self-help easy to find and use.
AI could help organise questions about IT security, but would not provide my security sign-off. Prototypes would use fictional information. Before introducing login and processing information about children and wellbeing, relevant specialists would review data collection, access and protection. Necessary data and any partner reporting would need to be defined before a pilot.
Value for families.
Value for professionals.
I would test a monthly membership and explore paying psychologists according to use of their content, with separate payment for reviewing others’ material. Visibility, future partnerships and referrals to consultations outside the platform could also be explored with professionals.
This is not yet a finished business model. Usage-based payments could reward consumption rather than useful help. Payment for professional review should not depend on whether the content is approved. These incentives need careful consideration before the model goes live.
I would track the cost of acquiring a paying family, membership duration and what remains after paying professionals and operating costs. A low acquisition cost only matters if the economics also work over time.
I would work on sales, partnerships and evidence alongside the product. Before each test, we would agree what would make us continue, change direction or stop.
Use is one signal.
Wellbeing is another.
I would distinguish between product activity, perceived help and changes in wellbeing. Relevant usage measures can show whether families start, complete and return. Questionnaires can contribute self-reported changes. Neither should stand alone as evidence that the product caused an improvement.
For an insurance partner, the central question is whether early self-help can reduce a later need for more extensive support, including while a family waits for other help. This would be a hypothesis to investigate with an appropriate evaluation design, not an already proven sales claim.
I would therefore agree measurement and value criteria early with a partner. Activity alone is insufficient, and more time in the app is not automatically a better outcome.
I would take responsibility
for the whole business.
I would use AI to reduce dependence on large development projects and make it easier to act on what we learn. At the same time, I would bring in more professionals earlier, test willingness to pay and make the value clear to families and partners.
AI changes my ability to build and test. My experience with MY INNER ME changes what I would prioritise: a more focused start, a sustainable business and clear human responsibility for quality and security.