Skills / Build It Right / Architecture & master data
Build It Right

Data & AI Architecture

Assesses the estate you have, designs the pattern you need, and reviews plans before they are funded.

Private skill · shared by invitation · How access works

Illustrative output of Data & AI Architecture: Assesses the estate you have, designs the pattern you need, and reviews plans before they are funded.
Illustrative output with sample data.
The challenge

The wrong pattern takes years to unwind.

Pattern by fashion

Mesh, fabric or lakehouse gets chosen because it is new, not because it fits.

AI bolted on

Vector stores and agents land beside the platform with no governance.

Decisions nobody wrote down

Two years later nobody remembers why the platform looks the way it does.

Reviews after the fact

Architecture is critiqued once the money is already spent.

How it works

From first question to finished output.

Assess

A 35-question assessment of the current estate, scored 0–100.

Design

Pick across EDW, Lake, Lakehouse, Hub, Mesh and Fabric from workload, team and governance maturity.

Design for AI

RAG versus fine-tuning, feature and vector layers, and governed agents.

Review

Pressure-test any proposal against a named anti-pattern catalogue, and record the decision.

What you get

  • 35-question assessment scored 0–100
  • Pattern decision across EDW, Lake, Lakehouse, Hub, Mesh and Fabric
  • Full AI stack design and Architecture Decision Records

What goes in

Current estate description, workload mix, team size, governance maturity, AI ambition.

Time to answer

As a consulting engagement14 weeks3 people · 42 consultant-weeks
Running the skill2–3 hours per modeone sitting, start to finished output

What it still needs from you

The political question of which team loses their preferred platform.

Also in Build It Right

See where this fits in the sequence →

See Data & AI Architecture in action.

Thirty minutes with Harveer, on your own data problem.