AI for your role

AI for Analytics Engineers

Let AI scaffold the models so you can obsess over the tests that make data trustworthy.

Get the Analytics Engineer brief
The shift

How AI is changing the Analytics Engineer role

In 2026, AI copilots scaffold dbt models, write SQL transformations, and draft tests and documentation faster than any Analytics Engineer could by hand. That makes the role more valuable, not less: value concentrates in modeling data so downstream teams can trust it, designing tests that catch silent breakage, and shaping the semantic layer the whole company reasons on.

What AI can take off your plate

  • Writing dbt models and SQL transformations from a spec
  • Generating schema tests and data-quality assertions
  • Drafting model documentation and lineage descriptions
  • Refactoring repetitive transformation logic
  • Explaining a failing test or a data-quality incident

What stays distinctly human

  • Modeling data so downstream trust is earned, not assumed
  • Designing tests that catch silent data breakage
  • Owning metric definitions in the semantic layer
  • Making grain and dimensional-model decisions
  • Balancing warehouse cost, freshness, and performance
Tools

Five AI tools for Analytics Engineers

dbt (with dbt Copilot)
Generates models, tests, and documentation from natural language and centralizes metrics in the semantic layer.
Try it →
ChatGPT
Writes and refactors SQL, drafts dbt tests, and explains a gnarly window function or a failing assertion.
Try it →
Claude
Reviews a transformation for grain and join bugs, proposes edge-case tests, and drafts clear model docs from your code.
Try it →
Cursor
An AI-native editor that understands your dbt project and helps write, refactor, and debug models across files.
Try it →
Datafold
AI-assisted data-diffing and column-level lineage that shows exactly what a change to a model will break downstream.
Try it →
Prompts

Five prompts to try today

Paste these into Claude or ChatGPT and replace the bracketed parts with your own details.

1. Draft a dbt model
I need a dbt model for [purpose] from source tables [describe]. Write the SQL following [conventions], set the right grain, and suggest the schema tests I should add.
2. Design the tests
Here's a dbt model: [paste]. What silent failures could slip through — null spikes, grain changes, broken joins, duplicates? Write dbt tests to catch each.
3. Debug a data-quality incident
A downstream metric jumped [describe]. Here are the relevant models [paste]. Walk through the likely causes upstream and how to isolate which transformation introduced it.
4. Optimize warehouse cost
This model is expensive [paste SQL and describe run frequency]. Suggest incremental strategies, materialization changes, or refactors to cut cost without hurting freshness.
5. Define a semantic metric
We want a governed definition of [metric] in the semantic layer. Propose the definition, the dimensions it should be sliceable by, and the tests that keep it honest.
The playbook

Every AI play for Analytics Engineers

Your full AI playbook for your role — updated every week. Tap any card for a step-by-step walkthrough and examples.

✦  New AI plays are added every week — and go straight to subscribers in their morning brief. Skip the scrolling and get yours delivered free. Get my free brief →
Loading the library…

A day in your inbox

This is the kind of brief a Analytics Engineer gets, every weekday morning.
Monday morning
✦ Personalized for: Analytics Engineer
Data PlaybookAd-hoc data pull
Get answers from a messy CSV without writing a single query

The move for when a stakeholder wants numbers now and you do not want to spin up a notebook. Free tier, plain English in, charts out.

Julius AI FREE  a free tier that analyzes spreadsheets and data in plain English, with charts

1

Go to julius.ai (free account), start a new chat, and upload the raw export the stakeholder sent you, for example the sales_export.csv.

2

Ask it in one line, so it cleans and computes in one pass:

Using [sales_export.csv], group revenue by [region] and [month], drop any rows where [amount] is blank or negative, and show me the top 3 regions by total revenue as a bar chart.
3

Then pressure-test the result before you send it up:

How many rows did you drop and why? Show me 5 example rows you excluded so I can confirm the logic.

You answer a same-day request in minutes with a chart and a clean audit trail, instead of hand-writing SQL against a file nobody has profiled yet.

Your role, all in one place
  
Tools, prompts & tricks
Your full library, one tap away.
›
  
Your playbook
Every entry, building each week.
›
  
How AI is changing your role
Where your work is heading.
›

You’re subscribed as Analytics Engineer.  ·  Update your roles  ·  Manage preferences  ·  Unsubscribe
The Morning Current · Powered by Atomic Media Group, LLC

Get the Analytics Engineer brief

One AI play, built for your role, every weekday morning. Free.

You’re in! We just emailed your first brief — it should land in a minute. Add brief@themorningcurrent.com to your contacts so it never hits spam.
Free forever. Unsubscribe anytime. We use your role only to personalize your brief.