AI Is Your New Peer Programmer
Chakkaradeep Chandran
I lead AI product. I still build.
Your coding agent will do whatever you ask. That is the problem.
Ask a coding agent for a small change and you get back something that looks finished: a tidy set of edits, a confident summary, and every check passing. All three can be true while the change is still wrong for the product around it. Every requirement you leave vague is a decision handed quietly to the model, and you usually find out which ones only later, when something breaks in a way nobody predicted.
Much of the advice about these tools is about getting them to produce more. This session is about the other half of the job: directing the work, then judging what comes back.
The method has three phases. Ask establishes the context the agent cannot infer, and makes it show that it understands how the thing works today before changing anything. Plan turns product intent into something specific enough for an engineer to argue with. Test asks what evidence would actually prove the change does what you asked for.
What you’ll learn
You stop taking the agent's word for it, and know what to check instead.
The three phases, and the decision point that sits between each one.
What you decide before the agent starts: what changes for the user, and what you are deliberately leaving out.
A live demo in a widely used open-source product, not a sample app built for teaching.
What to look for when it explains how something works today, and whether it can show you where.
How to tell a plan that holds up when you argue with it from one that only reads well.
A checkpoint in the demo where we stop and grade its answer together, before anything gets built.
The prompts that make an agent argue against its own plan, and what a real revision looks like.
The plan changing live, and the decision that only surfaced because somebody pushed back on it.
Then the build, a test that fails before it passes, and what makes a claim of done believable.
Workshop agenda
The decisions you hand over without noticing
If you don't say it, the agent decides it. Scope, what happens when things fail, what gets tested, how it ships. We also look at what you own and what your engineers own.
What you need before any of this is possible
Which tool, which mode, and how much access to the code you really need. How to ask your engineering lead for it, and what you can still do if they say no.
Ask: understanding the system before changing it
The decisions that are yours before work starts. How to make the agent explain how something works today, and how to tell a real explanation from a convincing one.
Plan: writing something an engineer can argue with
What makes a plan someone can push back on, instead of one that just reads well. How to attack it yourself before anyone builds. The written exercise sits here.
Test and verify: what would actually prove this works
What proof would break if you removed the change. Then the different question: is what got built the thing you agreed to?
Where this goes wrong, and what to do next
The ways this goes wrong, and how to spot each one. How to make a change easy for someone else to review. Then pick one small thing to try yourself.
Learn directly from Chakkaradeep
Chakkaradeep Chandran
AI Product Leader @ GoDaddy | ex-Microsoft
Who this workshop is for
Product Managers who can read code but do not write it. You have views on how things get built and want them backed by more than instinct.
Product Managers already using a coding agent. You have shipped something with AI help and had no real way to check whether it was right.
Product leaders deciding how their teams use these tools. You want a position formed by watching the work, not by reading about it.
What's included
Live sessions
Learn directly from Chakkaradeep Chandran in a real-time, interactive format.
Lifetime access
Go back to course content and recordings whenever you need to.
Community of peers
Stay accountable and share insights with like-minded professionals.
Certificate of completion
Share your new skills with your employer or on LinkedIn.
Maven Guarantee
Your purchase is backed by the Maven Guarantee.
Frequently asked questions
Maven for Teams
Reimbursement
Get your company to pay
Everything L&D needs: email template, receipts, and certificate of completion.
Get reimbursedTeam discount
Learn with your teammates
Save 20%+ when 2 or more teammates enroll in the same cohort.
Save 20%+ with a teamPrivate cohort
Run a cohort for your org
A dedicated cohort with a custom schedule and curriculum, tailored to your team.
Book a private cohort$250
USD
12:30–2:30pm EDT