Clinician-Directed. Culturally Grounded.
A Clinician’s Voice Shaping a Live AI Layer.
64 Companions. Every response clinician-directed. The voice of a 30-year clinical practice, shaping a live AI layer through an authored system prompt.
The digital mental health market is projected to reach $17.5 billion by 2030. Nearly every major player relies on generative AI running in real time — producing novel responses on the fly from training data that is overwhelmingly Western, white-normative, and anglophone.
This creates a structural vulnerability: the intelligence generating those responses has no lived understanding of the communities it claims to serve.
| User Experience | AI Interpretation | ENRICH Interpretation |
|---|---|---|
| Jamaican woman: “My mother stays in my business” | Enmeshment, boundary violation | Relational closeness normative in collectivist families |
| Nigerian man holding grief for extended family | Emotional avoidance, suppression | Collective grief-holding as cultural practice |
| Dominican teen consulting grandmother before decisions | Codependence, lack of autonomy | Intergenerational wisdom-seeking; healthy relational structure |
| Indian woman prioritizing family honor | Self-sacrifice, people-pleasing | Navigation of duty and selfhood within a collectivist moral framework |
The Intelligence That Heals You Was Never Artificial.
ENRICH’s response is not to reject technology. It is to ensure that the intelligence — the clinical judgment, the cultural understanding, the therapeutic intentionality — comes from a human clinician, not from a statistical model.
Every ENRICH companion is built on our proprietary, clinician-authored architecture — nearly 30 years of clinical practice encoded into a system prompt that directs a live AI layer in real time. The model does not invent the clinical voice. It carries the voice that Sherry Jerimie spent 30 years authoring. No conversation history is stored on a server after your session.
AI is the delivery layer. The intelligence itself — the decisions about what to say, how to frame cultural experiences, which therapeutic approach serves this specific wound — comes from three decades of clinical practice with the Global Majority. That practice is encoded into the system prompt that shapes every response.
deployed companions — the largest library of culturally-responsive wellness tools in market
conversation history stored on a server after your session ends
years of clinical authorship that no algorithm can replicate
people in the Global Majority — the largest underserved market in digital wellness
Our Intelligence Is Earned, Not Engineered.
The clinical voice was authored over 30 years and encoded into the system prompt that shapes every response. The model carries the voice. The clinician owns the intelligence.
ENRICH begins with the cultural wisdom of the Global Majority. The culture is not the footnote. It is the foundation.
Your messages are sent to a live AI service to generate each response. No conversation history is stored on an ENRICH server after your session. Your stories are never used to train any model. Data sovereignty is a requirement, not a feature.
The capacity to understand what it means to carry your grandmother’s silence in your own throat. That understanding is earned, not computed.
For most of the world, distress lives in relationships, families, communities, and histories. ENRICH was built from that understanding.
“I’m fine” in a Caribbean family means something different than in a suburban American household. Our companions know the difference.
Every response can be reviewed, challenged, and improved. A human clinician stands behind every word — not a black-box algorithm.
| Dimension | Generic AI Wellness Tools | ENRICH |
|---|---|---|
| Intelligence Source | Generic models, no clinical direction | Clinician-directed, 30+ years practice |
| Cultural Framework | Western frames with cultural filters bolted on | Global Majority epistemology as foundation |
| Runtime AI | Generic chatbot, no clinical direction | Live AI layer directed by a clinician's authored system prompt |
| Creation Process | Fully automated, no clinician in the loop | Clinician-authored system prompt directs the model |
| Data Privacy | Conversations stored, sometimes used for training | No conversation history stored after session; never used for training |
| Target Population | General (Western-centered) | Global Majority (4 billion people) |
| Accountability | Black-box algorithm | Named clinician behind every response |
| Content Library | Single generic chatbot | 64 specialized companions |
This Intelligence Knows Your Grandmother’s Name.
We believe that intelligence in a therapeutic context has never been about generating novel sentences. It is about knowing which response this specific person needs in this specific moment.
A large language model can produce a grammatically perfect paragraph about intergenerational trauma. It cannot understand what it means to carry your grandmother’s silence in your own throat.
The clinician who directs ENRICH’s companions can. That understanding — earned, embodied, culturally situated — is what we mean by Sovereign Intelligence™.
AI is the delivery layer, not the architect. The frontier model that carries the voice does not know who will use the companion. It does not understand why a Trinidadian woman’s “I’m fine” means something different than the same words spoken in a suburban American household. But the clinician who authored the system prompt knows. That authored voice is what reaches you.
Sherry knows. That is Sovereign Intelligence™.
— Sherry Jerimie
Certified Psychoanalyst & Licensed Psychotherapist
Founder, ENRICH Global
Everything you want to know about Sovereign Intelligence™ and how ENRICH uses technology responsibly.
Yes — honestly and transparently. ENRICH companions use a live AI layer, directed in real time by a clinician-authored system prompt. The system prompt was written by Sherry Jerimie, a Licensed Psychotherapist and Certified Psychoanalyst with nearly 30 years of experience serving the Global Majority. It encodes how Sherry thinks, listens, and moves in a clinical room.
Here is the critical distinction: the AI is the delivery layer, not the intelligence. The model that carries the voice did not author the voice. The voice was authored by a clinician over 30 years and encoded into the prompt that shapes every response. The model is constrained by that authored voice. Without the system prompt, you would get a generic chatbot. With it, you get a clinical voice trained for nearly three decades on the wounds of the Global Majority.
This distinction matters: a black-box chatbot has no clinical accountability. ENRICH has a named clinician behind every word. If a response feels off, the system prompt that shaped it can be reviewed, refined, and redeployed. That is the difference between Sovereign Intelligence and a generic AI tool.
Because most AI wellness tools deploy a generic model with no clinical voice, no cultural specificity, and no clinician-authored constraints. When a Jamaican woman says “my mother stays in my business,” an unconstrained model interprets enmeshment and boundary violation. A clinician who understands Caribbean family culture recognizes relational closeness normative in collectivist families.
That gap — between statistical prediction and embodied clinical understanding — is precisely where cultural harm happens. ENRICH closes that gap by directing the AI with a system prompt authored by a clinician who has spent 30 years sitting with the Global Majority. The model still does the language work. The clinician does the clinical work. That is what makes the intelligence Sovereign.
Your messages are sent to a live AI service to generate each response — ENRICH cannot deliver real-time conversation without that. But here is what data sovereignty means in practice: no conversation history is stored on an ENRICH server after your session ends. Your reflections are not logged for analytics. Your stories are never used to train any model. Anchor person, capacity check, and ENRICH Identity stay on your device, not in our database.
The frontier AI providers we use bind themselves to enterprise data terms that prohibit training on customer inputs. We chose them specifically for that reason. When you close your session, the conversation is gone from our side. That is what data sovereignty means in practice for a system that uses live AI responsibly.
Sherry Jerimie — Certified Psychoanalyst and Licensed Psychotherapist with nearly three decades of clinical experience serving the Global Majority. Her practice spans Caribbean, African, South Asian, East Asian, MENA, and Latinx communities across the diaspora.
Every companion is shaped by a system prompt that Sherry authored. That system prompt is what directs the live AI layer at runtime. It encodes how Sherry would respond, what she would notice, how she would honor cultural specificity, and where she would refuse to over-promise. The model carries the voice. The voice carries the weight of thousands of hours in the consulting room and the ethical accountability of a licensed clinician standing behind her work.
Sovereign Intelligence™ is clinician-directed intelligence — a clinician’s 30 years of practice, encoded into a system prompt that directs a live AI layer in real time. It is “sovereign” because the intelligence belongs to the clinician and to the communities the clinician has spent decades serving, not to whichever tech company built the underlying model.
The architecture is honest about what each layer does. The frontier model provides the language capability. The clinician-authored system prompt provides the clinical voice, the cultural specificity, the safety perimeters, and the therapeutic intentionality. Without the system prompt, you would get a generic chatbot. With it, you get a clinical voice that has been refined for nearly three decades on the wounds of the Global Majority.
Adding a “cultural sensitivity filter” on top of an undirected model does not make it culturally responsive. It makes it politely shallow.
True cultural responsiveness requires understanding the epistemological frameworks of the communities being served — not just their demographics, but their ways of knowing, healing, and relating. AI can be a powerful tool when directed by someone who holds that understanding. Left undirected in the therapeutic space, the risk of pathologizing normal cultural behavior remains too great.
ENRICH’s approach: use AI as the delivery layer, direct it with a clinician-authored system prompt encoding nearly 30 years of practice, build safety perimeters into the architecture, and route cultural content by the user’s ENRICH Identity. The technology serves the clinical voice. The clinical voice serves the community.
No. ENRICH companions are psychoeducational tools — they are therapeutic, not therapy. They help you understand yourself, your patterns, your cultural inheritance, and your possibilities. They provide the language and framework for self-understanding that can complement therapy, prepare you for therapy, or support you when therapy is not accessible.
Every companion includes crisis resources, daily limits to prevent emotional overwhelm, and clear boundaries about what a companion can and cannot do. If you are in crisis, ENRICH will always direct you to appropriate clinical support.
Because the system prompt that shapes every response is auditable. The clinical voice is not an opaque algorithm: it is a written document, authored by a licensed clinician, that you can read about and that we continuously refine based on what we learn in practice.
This is one of the Seven Principles of Sovereign Intelligence™: Auditable and Accountable. If a response feels wrong to you, that feedback goes to the clinician who authored the system prompt. The prompt can be reviewed, refined, and redeployed. There is a named human being accountable for the voice. That is something a generic chatbot cannot offer.
64 wellness companions. Clinician-directed intelligence. Every response carrying the weight of nearly three decades of practice with the Global Majority.
Meet Our Companions