End-to-End Tutorial

This tutorial walks through the full lifecycle of a data product: from writing manifests to querying results in a serving endpoint. You will create a “customer-orders” aggregate product that joins cleaned customer data with order events, deploys it to the platform, runs it, and queries the output.

flowchart LR
    S1[1. Init Project] --> S2[2. Write Manifests]
    S2 --> S3[3. Validate]
    S3 --> S4[4. Register]
    S4 --> S5[5. Deploy]
    S5 --> S6[6. Trigger Run]
    S6 --> S7[7. Monitor]
    S7 --> S8[8. Query Results]
    S8 --> S9[9. Check Quality]
    S9 --> S10[10. View in Catalog]

Prerequisites

Before starting, ensure you have:

• The akili CLI installed and configured with API credentials
• Access to a tenant (ask your platform admin for tenant access)
• At least one connection registered (for source products) or existing upstream products (for aggregates)

# Verify CLI is configured
akili whoami

You should see your username, tenant, and role.

Note

This tutorial creates an aggregate product that depends on two upstream products: cleaned-customers and cleaned-orders. In a real deployment, those would already exist. The tutorial includes complete manifests so you can follow along regardless.

---

The Tutorial Product

We will build customer-orders — an aggregate product that:

• Joins customer data with order events
• Computes per-customer order metrics
• Serves data via lookup (Portal), analytics (dashboards), and realtime (live tiles)
• Enforces quality gates on data freshness, completeness, and referential integrity
• Classifies the output as internal

Overview of the 10 steps:

Step	Action	Tool
1	Initialize project	akili init
2	Write all 6 manifests	Text editor
3	Validate manifests	akili validate
4	Register with platform	akili product create
5	Deploy to execution engine	akili deploy
6	Trigger execution	akili run trigger
7	Monitor progress	akili run status
8	Query results	REST API / SQL
9	Check quality scores	akili product quality
10	View in catalog	Portal / akili search

---

Continue Reading

• Step 1: Declare — Initialize and write the 6 manifest files
• Step 2: Deploy — Validate, register, deploy, and trigger execution
• Step 3: Query — Monitor, query results, check quality, and discover