How EQ Works

From the first text message to the session prep summary on your dashboard — here is exactly how EQ works, end to end.

Step 1: Patient Enrollment

A therapist registers on EQ at my-eq.com and adds their patients to their dashboard. Each patient is linked to a dedicated Twilio phone number provisioned through EQ’s infrastructure. No app download is required. No account creation is needed for the patient.

Step 2: The Patient Receives a Text

EQ initiates contact with the patient via SMS. The message arrives from a consistent phone number so patients recognize it over time. The patient responds naturally, in their own words, as if texting a person.

Step 3: The AI Conversation Engine

Each incoming message is received by EQ’s serverless processing layer, built on AWS Lambda. The message is passed through a LangChain pipeline backed by OpenAI’s large language models. This pipeline does two things:

  1. It generates a warm, therapeutically-appropriate response to continue the conversation
  2. It performs emotional extraction — identifying emotions present in the message, assessing their intensity, and categorizing them against a structured taxonomy

The response is delivered back to the patient via Twilio within seconds. The emotional data is stored persistently.

Step 4: Emotional Data Accumulates

Over days and weeks, EQ builds a longitudinal emotional profile for each patient. Every message is a data point. The system tracks not just what emotions appear, but how they evolve across conversations — identifying patterns like recurring guilt themes, increasing anxiety signals, or a shift toward gratitude and stability.

Step 5: The Therapist Dashboard

Before each session, the therapist logs into their EQ dashboard. The dashboard is powered by a Go-based REST API served from AWS ECS Fargate. For each patient, the therapist sees:

  • A summary of recent emotional conversations
  • Dominant emotional themes and their intensity rankings
  • A session prep prompt generated from the patient’s current emotional state
  • The full conversation history for context

The session prep prompt is not templated. It is generated fresh using the patient’s actual recent data, so every session brief is specific to that patient’s current emotional reality.

Step 6: The Session

The therapist walks into the session already oriented. Instead of spending the first 10–15 minutes of a session asking “how was your week,” they can open with a specific, informed question grounded in what the patient has actually been experiencing. The conversation can go deeper, faster.

The Full Stack

For those who want the technical picture:

  • Patient SMS layer: AWS Lambda + Twilio (Go runtime)
  • AI pipeline: LangChain + OpenAI (emotional extraction and conversation generation)
  • Therapist API: Go + Fiber framework, JWT authentication
  • Data persistence: Managed relational database on AWS ECS Fargate
  • Infrastructure: Terraform (infrastructure as code, fully reproducible)
  • Frontend: React-based therapist dashboard at my-eq.com

EQ is not a monolith. It is a purpose-built, cloud-native system designed from day one for clinical use cases where reliability, privacy, and data fidelity are non-negotiable.