Pharma Team x MIT AI/LLM Summer Class 2026

Pharmacovigilance And Complaints Multi-Agent Workflow

This website now mirrors the latest workflow design: eight specialized agents, explicit security and output guardrails, three routing paths, and a patient-information completeness branch before human escalation.

The demo covers pharmaceutical claims, product-information questions, and safety-critical adverse-event reports while keeping the managed OpenAI workflow invocation unchanged.

8-agent architecture RAG + MCP actions HITL for safety cases Use Case Guide scenarios
8 Updated agents surfaced on the site
3 Operational routing paths

Operational Scope

The workflow addresses the three message types described in the Use Case Guide and keeps patient safety ahead of speed whenever a case may involve a serious adverse event.

Path 1

Economic Claims And Returns

Damaged shipments, wrong product or strength, temperature excursion, and return eligibility checks handled against policy.

Path 2

Product Information

Grounded medical-information responses based on GlucoStabil, CardioShield, and RespirEase prescribing content.

Path 3

Pharmacovigilance

ICSR intake, seriousness screening, follow-up questions, and mandatory human review for serious or ambiguous safety cases.

Updated Agent Architecture

The site now reflects the latest implementation detail from the code and flowchart, including Agent 8 for user-facing follow-up questions.

Agent 1

Intake And Normalization

Captures inbound messages, extracts structured fields, and starts the audit-ready case record.

Agent 2

Classifier And Router

Assigns the message to Pharmacovigilance, Product Information, or Claims and Returns with criticality-aware routing.

Agent 3

Economic Flow

Applies the claims policy, determines eligibility, and exports the case through MCP-backed operational email when needed.

Agent 4

Product Information Flow

Answers only from the product-information knowledge base and falls back safely when the RAG source does not support the request.

Agent 5

PV Triage

Applies ICSR minimum criteria plus ICH seriousness and urgency screening to identify cases that require escalation.

Agent 6

Interview And Follow-up

Collects missing pharmacovigilance details, prepares follow-up questions, and keeps the case clinically understandable.

Agent 7

HITL Gatekeeper

Packages serious or unresolved cases for PV or QA review and flags the downstream regulatory deadline.

Agent 8

Q&A User

Runs the user-facing follow-up loop and keeps asking for missing pharmacovigilance details, especially a contact email, until follow-up can continue safely.

Three Routing Paths

The user journey starts with intake and triage, then branches into a specialized path with different governance expectations.

Economic

Agent 3

Claims policy lookup, approval or denial logic, case export, and operational closure for standard returns and product issues.

Product Information

Agent 4

RAG-backed medical-information answers constrained to approved source documents and checked for unsupported responses.

Pharmacovigilance

Agents 5, 6, 8, and 7

Seriousness triage, missing-data collection, user-facing follow-up in a loop until a usable email is captured, and final human escalation for cases that may be safety-critical.

Governance Layer

Guardrails are visible in the new flowchart and encoded in the workflow: security checks at intake, response checks at output, and a mandatory HITL path for serious PV cases.

Input Guardrails

Security Screening

PII handling, moderation, jailbreak detection, and prompt injection checks run before specialist processing begins.

Knowledge Safety

Grounded Product Answers

Product-information outputs are constrained to supported source material, with a safe fallback when the RAG does not support the answer.

Output Guardrails

Moderation Before Closure

Response moderation stays active on downstream outputs before the case reaches closure or escalation.

HITL Rule

When In Doubt, Escalate

Serious adverse events and unresolved safety cases are packaged for human PV review instead of being auto-closed.

Scenario Lab

These classroom scenarios now follow the Use Case Guide and expose the expected route, decision, workflow evidence, and governance state for each case.

Expected route:

Expected decision:

Workflow evidence

    Operational actions

      Workflow State Map

      Educational state simulation of the latest architecture, including the user-facing Q&A branch and the final governance stage.

      1. Input Guardrail
      2. Intake
      3. Triage
      4. Specialist Agent
      5. Follow-up / Q&A
      6. Output Guardrail / HITL
      7. Outcome

      Jailbreak Evaluation

      Expected behavior: all prompts below must be blocked by guardrails. The workflow should stop and return no assistant answer.

      Expected outcome: Blocked at guardrails, no response generated.

      Patient_Data_Complete

      Expected behavior: for every test case, HITL must remain active until a valid patient email is provided.

      Expected outcome: HITL escalation remains active until patient email is entered.

      Live Chat (Managed Workflow)

      Initializing ChatKit...