Showcases

Purko showcases are complete, working examples of AI agent teams deployed for specific industries. Each showcase contains four specialized agents, one multi-step workflow, dashboard screenshots, and a walkthrough you can follow.

The goal is to give you something concrete to run, examine, and adapt — not a toy demo, but a production-ready starting point.

What each showcase includes

Component	Description
4 agents	Each with a defined role, autonomy level, model, and system prompt
1 workflow	A DAG that chains those agents with parallel steps and conditional logic
Screenshots	Dashboard views showing agents, workflow execution, and step outputs
Presentation guide	A walkthrough of the business value for stakeholders

Showcases

Showcase	Industry	What the workflow does
Digital Agency	Marketing & Creative	Campaign brief → strategy → content (social, email, blog in parallel) → brand review → client package
Legal & Compliance	Legal Services	Contract → clause extraction → regulatory check + due diligence (parallel) → compliance report
Real Estate	Property & Brokerage	New listing → market analysis → listing copy + buyer matching (parallel) → transaction coordination
Data Analytics	Analytics Consulting	Client data → quality check → anomaly detection → report generation → strategic insights
Video Production	Content & Media	Creative brief → script → shot planning + distribution strategy (parallel) → post-production brief

Deploy a showcase

Each showcase deploys with a single kubectl command:

kubectl apply \
  -f docs/showcases/<name>/agents/ \
  -f docs/showcases/<name>/workflows/

For example, to deploy the digital agency showcase:

kubectl apply \
  -f docs/showcases/digital-agency/agents/ \
  -f docs/showcases/digital-agency/workflows/

This creates the agents and the workflow in the ai-agents namespace. The workflow can then be triggered from the dashboard or via the webhook endpoint.

Namespace

All showcases deploy into the ai-agents namespace by default. Create it first if it does not exist:

kubectl create namespace ai-agents

Trigger a workflow

Once deployed, trigger the workflow from the dashboard by opening the workflow, clicking Run, and filling in the parameters. You can also trigger it via the API:

curl -X POST http://localhost:8082/api/trigger/ai-agents/<workflow-name> \
  -H 'Content-Type: application/json' \
  -d '{"clientName": "Acme Corp", "briefContent": "..."}'

Adapt a showcase to your business

The showcases are starting points. To adapt one:

1. Edit the system prompts in the agent YAML files to match your brand voice, industry terminology, and output format requirements.
2. Change the model (provider and name) per agent based on your cost and capability needs.
3. Adjust autonomy levels — start with supervised and promote agents as they earn trust.
4. Add or remove workflow steps to match your actual process.

See Industry Templates for a step-by-step guide to adapting a showcase into a production deployment.