AI Data Analyst Blueprint

A cohort program for building and operating agentic AI systems that do data analysis work — and for being the person who can tell when they're wrong.

If you haven't read it yet, the complete method is published free, with working templates. That page is the honest test of whether this is for you: if you can build the system from it on your own, do that and save your money. This program exists for the part reading can't give you — building it against your own data, with feedback, alongside people doing the same thing.

What you'll be able to do

By the end, you should be able to:

Build an agentic AI system that connects to a real data warehouse, writes and executes SQL, and returns results with the query attached
Write the schema context and business-rule documentation that makes the difference between a demo and something you'd trust
Design QA loops that catch the failure mode that matters — a query that runs clean and returns a plausible, wrong number
Apply the same agentic patterns outside analytics, to other repetitive work you do on a computer
Explain and demonstrate what you built, in interviews or to a client, with a working prototype rather than a certificate

These are capabilities, not outcomes. What you do with them is up to you and to circumstances neither of us controls — see what this program does not promise.

What's inside

Phase 1 Agentic AI foundations

How agents actually work, and how to direct them: instruction files, tool and API access, controlling scope, managing cost, and the failure modes that appear once an AI can take actions rather than just answer questions. Worked examples beyond analytics, because the patterns transfer.

Phase 2 AI for data analytics

The full build from the free method page, done properly and slowly — against your data if you have warehouse access, or a public dataset if you don't. Environment, architecture, the agent instruction file, schema context, the build prompt, QA, deployment. Then the workflows the article doesn't cover: scheduled reporting, multi-source joins, and handing output to non-technical colleagues.

Phase 3 Putting it to work

Presenting the skill honestly and effectively — on a résumé, on LinkedIn, in interviews, or in a client proposal. How to scope and price this as freelance work if that's your route, including what ongoing maintenance actually involves. How to introduce it inside an organization without alarming the people whose jobs touch the same work.

Support

Live calls on a published schedule — Q&A and working sessions where we build things together. Recorded if you can't attend.
A private community where you can ask questions between calls. I'm in it and I answer.

Software built for this program

Two applications I built. They come with the program and aren't sold separately.

AI Job Interview Coach

Upload a job description and your résumé, and it runs a realistic interview for that specific role — questions drawn from the actual posting, not generic ones — then gives you detailed feedback at the end. Practice as many times as you want, on as many roles as you want.

AI Chris

A chatbot trained on my training material and this program's content, so you can get answers at two in the morning when I'm asleep. It's an AI, not me: good for "how should the context file handle this?" questions, less good for judgment calls about your specific warehouse. Those are what the live calls are for.

Also included

Data Analyst Academy — the complete curriculum from my previous mentorship, covering Excel, SQL and Python fundamentals. Included at no extra charge for anyone who needs the human-side skills first. You cannot check an AI's SQL if you can't read SQL. This previously sold for $4,800 as a standalone program; it is no longer sold separately at any price.
The AI Automation Business Package — training on building AI automations for small businesses as a freelancer: what to offer, how to price it, how to find clients, and how to deliver the work once you've won it.
Résumé & LinkedIn AI prompts for describing AI capability without overstating it.
Expert's guide to finding remote jobs — where remote analytics roles are actually posted and how to approach them.
Agentic AI Certification, issued by Data Analyst Academy, attesting that you successfully completed the main project — a working AI data analyst app. See the note under "who it isn't for" for what that certificate is and isn't.

Format and time commitment

Delivery	Recorded core curriculum plus live calls on a published schedule. All live sessions are recorded.
Length	12 weeks. You keep access afterwards, so you can go slower if you need to.
Time needed	Plan on 4–6 hours a week to actually build things. You can watch it faster than that; you won't learn it faster than that.
Access	Lifetime access to recordings and materials.
Other costs	Roughly $20–$100/month in AI API usage while you're building, depending on how much you run it. Hosting has a free tier. These are paid to third parties, not to me.

Prerequisites

You need to be comfortable running a command someone gives you in a terminal and reading the error if it fails. You do not need to know how to write application code — the agent does that.

You do need to be able to read a SQL query and judge whether it's correct, or be willing to learn that here. This is load-bearing and there is no way around it. An AI data analyst you can't check is a liability, not an asset. If you're starting from zero, the included Data Analyst Academy curriculum covers it, but budget real time for that.

Access to a SQL warehouse at work makes the program considerably more valuable, because you'll build against real data with real quirks. If you don't have that, we'll use public datasets — you'll learn the method, but you'll miss some of the messiness that makes the skill stick.

What this program does not promise

Read this part

The first cohort ran June to September 2026 and has just finished. What I don't have yet is outcome data — whether the program led anyone to a job, a raise, or paid client work — because those take longer to materialize than the program does. When there are outcomes, they'll be published here with names and dates. Treat any new program showing you impressive placement numbers with suspicion; there hasn't been time for them to exist.

Specifically, I am not promising that you will get a job, get a raise, keep your job through a layoff, land clients, or earn any particular amount. I don't control the job market, your employer, or how much work you put in. Anyone who promises you those things is either guessing or lying.

What I am committing to is the content and the support described on this page: the curriculum, the live calls, the community, and my answers to your questions.

Two further notes. This is a fast-moving field and some specifics will date — tools change, model behavior changes, best practice changes. I update the material, but assume you're learning to adapt rather than memorizing steps. And my previous program, CareerHacker Data Analyst Mentorship, is retired; you can see what it was and who came through it there. That is evidence about me as a teacher. It is not evidence about this program's results.

What cohort one says so far

The first cohort ran from June to September 2026 and has just finished. These messages were sent by students during the program, published with their permission. Surnames shortened at their request.

"This course was awesome. I have learned to do some AI coding that I thought would be hard but it was so easy with Chris coaching on an easy step by step gradient. The ability to code with AI has now made my job so much easier. I developed an app that I thought was impossible to do. Totally blown away."

"My late nights at work are now because I'm integrating Claude Cowork and I'm absolutely killing it because of your course."

"Built a project that takes exports from my fitness tracking and nutrition apps as well as my digital scale, and tweaks my meal plan and workout programming week to week. Continuing to automate a lot of my tedious tasks at work."

Worth being precise about what these do and don't show. Both describe agentic AI work in general rather than the data analyst build specifically. They're evidence that the agentic AI teaching works. They are not evidence about analytics outcomes or about jobs, and I won't present them as such. When there are outcomes from the analytics build, they'll be published here with names and dates.

Who this is for

Working analysts and data scientists who want to do the mechanical part of the job far faster, and to understand what's coming
People moving into analytics who want a differentiator beyond the standard Excel/SQL/Python résumé
People in other computer-based roles who want the agentic AI patterns, with analytics as the worked example
Freelancers and consultants who want a concrete, high-value service to offer
Business owners who'd rather understand their own data without hiring for it

Who it isn't for

Anyone looking for passive income, or a system that runs without them. This is a skill, and operating it is the job.
Anyone unwilling to read a SQL query. I'm repeating it because it's the single most common reason someone would waste their money here.
Anyone who needs an accredited qualification. You'll finish with working software you built and can demonstrate, plus a certification issued by Data Analyst Academy attesting that you completed the main project. That certificate comes from my own company, not from an accrediting body — it's a record that you built the thing, and it shouldn't be presented as an industry credential. In practice the working app is the more persuasive artifact anyway.

Enrollment is open. The price is shown on the checkout page — no sales call, no application step.

See pricing and enroll

The guarantee

Complete the program — do the projects and finish the material — and if you're not 100% satisfied it was worth every penny, email me. I'll refund you in full, plus send you a $100 Amazon gift card, to say sorry for wasting your time.

To qualify: complete all three phases and build a working AI data analyst against your own warehouse or a public dataset. Email me within 90 days of joining.

Questions

What does it cost?: The price is on the checkout page. There's no sales call and no application step — you can see it without talking to anyone.
What if it isn't for me?: Complete the program and build the project, and if you don't think it was worth what you paid, email me within 90 days of joining. Full refund, plus a $100 Amazon gift card for your time. The only condition is that you actually do the work — I can't make it worth your while if you don't.
Can't I just build this from your free article?: Yes, and if you can, you should. The article is complete and deliberately ungated. What it can't give you is someone looking at your specific warehouse, your specific broken query, and telling you what's wrong with it.
Do I need to be a data analyst already?: No. Roughly half of what makes this work is understanding data well enough to document it, and that's learnable. The included Data Analyst Academy curriculum exists for exactly this. But be realistic about the added time.
Will this work with my company's data warehouse?: Snowflake, BigQuery, Azure SQL, Redshift and Postgres all work. What matters more is whether you can get read-only credentials — that's an internal permissions conversation, and in a regulated environment it can be a slow one. Worth starting before you enroll.
Is this going to be obsolete in six months?: Specific tools will change. The architecture — separating direction from orchestration from execution, documenting context, verifying output — isn't tool-specific and has held up while the tooling churned. I'd rather tell you that than pretend otherwise.
What if I fall behind?: You keep access to everything, and the community doesn't close when the cohort ends. Plenty of people take longer than the nominal schedule. One thing to know: the guarantee window is 90 days from joining, and the program runs 12 weeks — so if you want to keep that option open, aim to finish roughly on schedule.
Who's teaching it?: Me, Chris Shupe, directly. I've worked as a data analyst since 2014 in permanent and contract roles, including at HSBC, Salesforce, American Express, and a major delivery marketplace. I'm self-taught. I built the system in the free article for my own contract work before I taught it to anyone.

Still deciding? Read the free method in full. It's the whole build, no email required. If it turns out to be useful on its own, that's a fine outcome.