complaints-review logo

complaints-review

Free

Summarise, categorise and draft regulator-ready responses from complaint / conduct files.

Visit Website
Related by shared tags and keywordsNamaste-AITornado Brainkimss-python-quickstart
generative-aiai-agentscustomer-serviceextract-structureports-and-adaptersaiai assistantautomation

Overview

complaints-review is a conduct-ops assistant for Complaints and Conduct teams at APAC banks. It takes a customer complaint or conduct file and produces four cited, audited, maker-checker artifacts: ComplaintSummary, Categorization, ConductFlag[], and DraftResponse. The summary covers issue, products, channel, timeline, and parties. Categorization assigns a category with root cause, severity, and regulatory relevance cited to policy. ConductFlag[] lists conduct red flags such as vulnerable customer, systemic issue, regulatory breach, and complaint-handling-deadline risk, each with severity, detail, and citations. DraftResponse is a regulator-ready or customer-ready draft grounded in policy and regulation; it is always a draft and is never sent by the system, because a human reviews and sends it.

The system is built as ports-and-adapters (hexagonal) on the Gemini Enterprise Agent Platform, pinned to asia-southeast1 for data residency. Complaint files carry customer PII, so the full R1 safety pipeline runs: redact, then guardrail screen on the way in and out, with everything audited to a write-once (WORM) store. Policy and regulatory guidance is retrieved from the shared enterprise-knowledge-base governed RAG.

This is an engineering-portfolio reference repo, and all complaint data is synthetic and fictional. It provides five deployment profiles: local offline, live with a local open-weight model, gcp managed stack, platform HTTP services, and onprem fail-fast placeholders. It exposes a CLI, HTTP API endpoints, and an A2A AgentCard for discovery.

Pricing

Free

Detailed plans have not been confirmed in our catalog. Check the official website for current limits and billing terms.

Visit Website

Prices and limits may change. Confirm the currency, billing period, seat minimum and usage caps on the official website.

Use Cases

  • Complaints and Conduct teams use complaints-review to produce a structured summary of a complaint file, including issue, products, channel, timeline, and parties.
  • Teams use complaints-review to categorize a complaint with a category, root cause, severity, and regulatory relevance cited to policy.
  • Teams use complaints-review to identify conduct red flags such as vulnerable customer, systemic issue, regulatory breach, and complaint-handling-deadline risk.
  • Teams use complaints-review to draft a regulator-ready or customer-ready response grounded in policy and regulation, with human review before sending.
  • Teams use complaints-review to run the full complaint review pipeline offline on a laptop without Google Cloud, an API key, or emulators.
  • Teams use complaints-review to deploy a production managed stack with Document AI, Agent Search, Gemini, Model Armor, DLP, Cloud Logging WORM, Cloud Trace, and Gen AI Evals.
  • Teams use complaints-review to integrate complaint review into applications through HTTP API endpoints for full review, summary, and draft response.
  • Teams use complaints-review to discover the agent through an A2A AgentCard.

Features

AI assistant

Natural language interface

Workflow automation

Reviews

No reviews yet. Be the first to share your experience!

Rankings & collections