Blueprints
Pre-built, production-ready AI agent systems you can deploy and customize in minutes.
Blueprints are fully designed, production-ready multi-agent systems built for common high-value enterprise use cases. Instead of starting from scratch in the Control Hub, you select a blueprint, answer a short setup wizard tailored to your company, and Qubitz deploys a customized version of the application directly to your AWS environment.
Each blueprint includes a complete agent architecture, pre-configured tools, knowledge base setup, and deployment pipeline. The setup wizard captures your company-specific requirements and the AI customizes the agents, prompts, and workflows accordingly before deploying.
How to Use a Blueprint
Go to Blueprints
Navigate to the Blueprints page from the main navigation. Browse available blueprints — each card shows the name, description, type (Industry or Horizontal), and number of agents.
Select a Blueprint
Click a blueprint card to open the detail view. Review the description and agent count, then click Use This Blueprint.
Complete the Setup Wizard
Answer the configuration questions specific to that blueprint. The wizard collects your company name, industry, channel preferences, integration targets, and any domain-specific requirements. Most wizards take 2-3 minutes to complete.
Blueprint is Customized & Deployed
Qubitz clones the blueprint, runs the AI customization pipeline on your answers, and redirects you to the Control Hub where you can see the customized agents and deploy to AWS.
Available Blueprints
| Blueprint | Type | Best For |
|---|---|---|
| Customer Support | Horizontal | Any industry needing omnichannel support |
| Intelligent Document Processing | Horizontal / Industry | Invoice, contract, medical record, loan processing |
| AI Voice Screening | Horizontal (HR) | Automated candidate interview at scale |
| Hyperpersonalisation | Horizontal (Marketing) | Customer intelligence and campaign personalization |
Customer Support
An omnichannel AI support agent that handles customer enquiries across multiple channels with full human escalation support.
| Detail | Info |
|---|---|
| Type | Horizontal |
| Channels | Web chat widget, Slack, WhatsApp, Phone (voice) |
| Escalation | Warm transfer to human agent (voice) or ticket queue (chat) |
| Integrations | Jira, Zendesk, Salesforce, ServiceNow, HubSpot, Shopify, Stripe |
| Languages | English, Hindi, Spanish, French, German, Arabic (and more) |
What it does: Handles customer intents like order status, billing, returns, troubleshooting, and account management. Draws from your existing knowledge sources (Help center, FAQs, product docs, policies, past tickets). Takes actions like creating tickets, processing refunds, looking up accounts, and resetting passwords. Escalates to a human when sentiment is negative, the customer asks for a person, or confidence is low.
Setup wizard covers: Company and industry, active channels, customer intent types, knowledge sources, CRM/ticketing integrations, allowed actions, escalation triggers, escalation destination (phone number for voice, team/queue for chat), tone, and languages.
Intelligent Document Processing (IDP)
An end-to-end pipeline that reads, understands, validates, and acts on your documents — automatically.
| Detail | Info |
|---|---|
| Type | Industry / Horizontal |
| Document Types | Invoices, contracts, medical records, loan applications, and custom types |
| Processing | Classify → Extract → Validate (3-way) → Review → Report |
| Human Review | Auto-triggered for low-confidence or flagged results |
| Output | Structured data, validation report, webhook delivery |
What it does: Accepts uploaded documents and runs them through a multi-stage pipeline. It classifies each document by type, extracts the fields that matter to your business, then puts every extraction through a three-way validation check — cross-referencing results across multiple AI systems to catch discrepancies a single model would miss. Results that fall below your confidence threshold are automatically routed for human review. Once approved, the final structured data and report are delivered to your downstream systems via webhook.
Setup wizard covers: Company name, document types to process, key fields to extract per type, validation rules, confidence thresholds, and webhook destination.
AI Voice Screening
An automated voice-based candidate screening system that conducts structured interviews and delivers assessed results.
| Detail | Info |
|---|---|
| Type | Horizontal (HR / Talent) |
| Channel | Voice (inbound/outbound via phone) |
| Interview Stages | Initial Screening, Technical Assessment, Behavioral, Cultural Fit |
| ATS Integrations | Greenhouse, Lever, BambooHR, Workday, SAP SuccessFactors, iCIMS, JazzHR |
| Output | Structured interview transcript + scored assessment |
What it does: Conducts voice-based candidate interviews using Amazon Nova Sonic. Evaluates candidates across your chosen criteria (technical skills, communication, problem solving, cultural fit, etc.). Draws from job descriptions, competency frameworks, and interview rubrics. Delivers structured assessments with scores and summaries for each candidate.
Setup wizard covers: Role types being hired, interview stages to include, evaluation criteria, knowledge sources (job descriptions, rubrics), ATS integrations, company culture description, hiring goal, and languages.
Hyperpersonalisation
A customer intelligence and campaign personalization pipeline that analyzes your customer data and produces targeted outreach strategies.
| Detail | Info |
|---|---|
| Type | Horizontal (Marketing / CRM) |
| Input | Customer dataset upload |
| Output | 8-sheet Excel workbook + multiple PDF reports |
| Intelligence Dimensions | WHO, WHAT, WHEN, WHERE, HOW MUCH, WHY, RISK |
What it does: Accepts your customer data, runs automated analysis across 7 intelligence dimensions to build detailed customer profiles, and produces personalized campaign strategies. Outputs include segmentation analysis, behavioral patterns, propensity scores, recommended messaging, optimal channel and timing, spend predictions, and churn risk scores — all packaged as a downloadable Excel workbook and PDF reports.
Setup wizard covers: No extended wizard -- data is uploaded directly and the pipeline configures itself based on the dataset structure.
Customizing a Blueprint After Deployment
Once a blueprint is deployed, it lives in your Control Hub like any other application. You can:
- Edit agent configurations, prompts, and models in the Agentic System tab
- Add or swap integrations and knowledge bases in Tool Integrations
- Test changes in the Test Bed before redeploying
- Use App Refiner to modify the generated frontend application
- Manage webhooks, API keys, and custom domains in Application Management