Senior Product Manager / Lead Product Manager

I just figure out what needs to be solved in what order, build relationships and get to work.

• 7 years as an Analyst (revenue, pricing, data science).
• 7 years in Product (travel ecommerce, media, fintech).
• Worked in growth, platform & core product roles.

I'm interested in GenAI, search, b2c.

George Nurijanian

PMs (especially juniors) should use Claude Code for the simple reasons that

- it will make you more comfortable with the terminal and git, erasing a lot of the anxiety about those "scary" tools engineers use
- you won't learn product sense, but you will get into the mind of shipping real stuff, not those Lovable prototypes that don't impress anyone. touch the metal (to a degree), don't play in the sandbox. so you'll be well ahead of your peers even if you haven't yet built a breakthrough product
- you'll be surprised how many great ideas you'll start to get, from automating workflows to trying new AI skills and building apps. if you were in a rut, it'll get you out of it
- and the cool thing is that THESE IDEAS WILL BE POSSIBLE. you won't just get an idea and be stuck dealing with implementation details! CC can do it all honestly
- install great product development plugins like compound-engineering (find more at prodmgmt.world/claude-code) -- they will help you think through product decisions, it's like pairing with an engineer who won't sneer at you for asking questions
- if you get access at work, omg even better. but honestly $100-200 for a magical genie in a bottle that can do anything on your computer is dirt cheap and it will get cheaper i think. you earn enough, stop being cheap
- if you bought into the anti-AI hype, try to apply inversion for a moment. what if these slop peddlers on X who are talking about AI all day... what if they are right? upside is uncapped, downside is minimal, this is a perfectly asymmetric bet. it's Pascal's wager in new clothes. it's a better bet to believe than to not believe.

Daily product management tips & tools • Get my Product Management System, stop worrying about work • See what’s inside ↓

Open article: Claude Code for Product Managers by prodmgmt.world

Junior PM: I feel like I'm doing PM wrong. Our company doesn't work like anything in the books.

Senior PM: Which books?

Junior PM: Cagan. Ries. All of them. They talk about empowered teams and continuous discovery. We have none of that.

Senior PM: And this makes you feel like you're failing.

Junior PM: Constantly. I keep trying to implement these frameworks but they never stick.

Senior PM: Have you asked yourself who those books were written for?

Junior PM: For product managers. Obviously.

Senior PM: No. Those books describe how things worked at specific companies, during specific eras, with specific leaders.

Junior PM: But they're best practices.

Senior PM: They're case studies dressed up as universal truths. 

Junior PM: So the books are wrong?

Senior PM: The books are accurate. For those companies, at that time. Your company is neither of those.

Junior PM: So what am I supposed to do? Ignore everything I've learned?

Senior PM: Read the book. Note the principle. Then look at your actual company.

Junior PM: Look for what?

Senior PM: What works here. Not what should work. Understand the problem, don't jump to the solution - same advice as you would get about solving your customer problems, just turned inwards.

Junior PM: That feels like giving up.

Senior PM: It's the opposite. You stop being a student applying theory. You become a scientist running experiments.

Junior PM: But how do I even know what works if I've never seen good product management?

Senior PM: Watch who ships. Who actually gets things out the door.

Junior PM: Ok, yes, I can think of a few examples.

Senior PM: There you go. Study them. What do they do differently?

Junior PM: They never wait for permission. They build small things, show people, and suddenly it's real.

Senior PM: You just described better product management than most books.

Junior PM: But I thought we needed proper discovery. User research. Evidence.

Senior PM: You need to understand constraints. Know why something is hard. Detect when someone is hand waving.

Junior PM: That's it?

Senior PM: That's more than most PMs ever learn. The books give you vocabulary. Your company gives you reality. Your job is to bridge them.

Junior PM: So I should stop feeling bad that we're not like the companies in the books?

Senior PM: Those companies mostly don't exist anymore. Or they've changed so much the authors wouldn't recognize them.

Junior PM: That's oddly comforting.

Senior PM: The PM influencers on LinkedIn describe scenarios that happened at maybe ten companies total. Everybody else is figuring it out as they go.

Junior PM: So what's the actual job then?

Senior PM: Observe what works at your company. Contrast it against the literature. Build your own hypothesis about how to ship things.

Junior PM: That sounds harder than just following a framework.

Senior PM: It is. But it's the only thing that actually works.

I lowkey predict product roadmaps will look nothing like they do today by 2028

- quarterly planning cycles will compress to monthly or less
- "features" disappear as units of work, replaced by outcome targets agents ship against continuously
- the roadmap itself becomes a live dashboard of experiments running, not a Gantt chart of what's coming
- PMs stop thinking in terms of narrow slices of the products/features they own and start shipping complete experiences
- orgs will compete for compute the way they once competed for engineering talent
- the quality of the specs will still matter, but more so for how well the agents can execute them
- product ops will merge into devx and will have access to agentic traces looking for automation opportunities

a new era is starting.

The PM plugin is telling

Ten Anthropic-verified plugins at launch. Product Management alongside Productivity and Marketing.

That's where Anthropic sees the highest-value enterprise use cases in 2026.

For PMs, this has `/write-spec` for PRDs, `/roadmap-update` for stakeholder comms, `/synthesize-research` for discovery synthesis, `/competitive-brief` for market intel, `/metrics-review` for product analytics.

I'll be testing this out to see how it stacks up against other alternatives shortly, but I already know this won't be as in-depth as my library of skills (https://lnkd.in/gC_dM6Gf)

No alternative text description for this image

Most PMs don't use Claude Code at all.

Meanwhile, there are PMs using it to run their entire product workflow - from strategy docs to engineering tickets.

Here are 6 ways to actually use Claude Code as a PM:

1. Prototyping (the obvious one)

Build working MVPs in hours instead of weeks of cross-functional coordination.

But this is where most PMs stop. They treat it like a fancy wireframe tool.

2. Product strategy docs

Point Claude Code at your user research, competitive analysis, and metrics.

Ask it to draft your strategy doc. It sees connections across documents that you miss when switching tabs.

The output needs editing. But you're editing instead of staring at blank pages.

3. Order of operations consulting

"I have these 12 things to do. What order makes sense?"

Claude Code can read your project context, see dependencies, and suggest sequencing that accounts for blockers you forgot about.

It's like having a principal PM on call to gut-check your priorities.

4. Research synthesis

15 user interviews sitting in transcripts.

Claude Code processes all of them in minutes, surfaces patterns, contradictions, and quotes you'd miss after interview fatigue sets in.

Do this while research is fresh - not 2 weeks later when you've forgotten half the context.

5. PRD drafting

A decent PRD draft in 10 minutes instead of 2 hours.

Give Claude Code your user problem, success metrics, and constraints. It generates the structure.

You spend time refining the thinking, not formatting sections and remembering what template to use.

6. Complete workflow automation

User research to PRD to stakeholder slides to engineering tickets.

One continuous session. Each output feeds the next.

This is where Claude Code stops being a tool and becomes a junior PM who never gets tired and has perfect memory.

The common thread across all of these: Claude Code works better when it has context.

Feed it your actual docs. Point it at your real files. Let it see the full picture.

That's when it stops being a chatbot and starts being useful.

i compiled 11 must-read articles on AI agents I have read and favorited in my Reader app

had to filter for the last 3 months, even that feels old-ish
heck, 50% is by Anthropic, go figure 🤷 

here's the full list:

1. https://lnkd.in/eRGB77qm

2. https://lnkd.in/e8X9rZvK

3. https://lnkd.in/efbEpPUa

4. https://lnkd.in/egTxPZbY

5. https://lnkd.in/eAcRsTJj

6. https://lnkd.in/efXKxcaS

7. https://lnkd.in/ey58piNy

8. https://lnkd.in/ex-MAeVp

9. https://lnkd.in/ektus5Fi

10. https://lnkd.in/ei46yMEd

11. https://lnkd.in/eR96xPMK

Open article: A Guide to Claude Code 2.0 and getting better at using coding agents by sankalp.bearblog.dev

A Guide to Claude Code 2.0 and getting better at using coding agents

I've been reviewing product opportunity selection decisions from product teams (including my past work) and I keep seeing the same mistake. Let me share what separates the best from the rest:

1. Most PMs kill their best opportunities before exploring them. They ask "how hard is this to build?" during opportunity selection. That question is poison.

2. When you estimate effort before committing to an opportunity, you're not evaluating the opportunity. You're evaluating the first solution that comes to mind. These are wildly different things.

3. RICE and ICE frameworks train you to weight effort heavily. This makes sense for solution prioritization. For opportunity selection, it introduces solution bias before you've explored the solution space.

4. A team I know rejected "help users find relevant content" because engineering estimated 6 months for a personalization engine. A different team committed to the same opportunity without asking about effort. They found a 2-week solution with manual curation that solved 80% of the problem.

5. The second team separated opportunity selection from solution selection. They committed to the customer problem first, then found creative ways to solve it cheaply.

6. Constraint-driven creativity only happens when you've already committed. "We must solve this with minimal eng resources" produces better solutions than "is this worth our eng resources?"

7. Another filter that catches hidden problems: every opportunity must be solvable independently without touching sibling opportunities. If you can't address opportunity A without also fixing B, you've smuggled in a solution assumption.

8. Most "opportunity statements" are actually themes or outcomes disguised as opportunities. Real opportunities are tied to a single moment in the customer journey. This prevents boil-the-ocean thinking.

9. Speed matters more than accuracy in early opportunity selection. Accept conservative mid-scores for missing data and flag the gaps. Getting to assumption testing in week 1 teaches you more than three weeks of "better" analysis.

10. Customer satisfaction should be deliberately under-weighted in your scoring. Pain alone doesn't make something worth doing. Strategic fit predicts long-term defensibility. High-pain, low-fit opportunities are traps.

11. Your first opportunity selection will be wrong. The goal is to learn why it's wrong, not to avoid being wrong. Perfectionism in opportunity selection is just analysis paralysis with a fancier name.

If you want a prompt that implements this exact framework with distinctness filters and scoring weights built in, I put together a collection of 180 AI prompts for every stage of product work:

I spent a day testing Anthropic's PM plugin through a complete workflow - research synthesis to PRD to stakeholder update. 
                                                         
The outputs look professional. They fail every reality test.                                                                                  
                                                                                                                                                
But that's not the point. The plugin gives executives permission to ask "why do we need so many PMs?"                                         
                                                                                                                                                
What individual PMs do about this matters more than plugin quality.                                                                           
                                                                                                                                                
Full test results:

https://lnkd.in/ecqq5Smk

Claude Code is for technical work:

- debugging across 6 files to track down a state management bug 
- refactoring a component that touches 12 different imports                                                       
- git operations when I need to see which files changed                                                              
- anything where I need the full codebase context                                                                  

Claude Cowork is for admin work:   

- filling out forms in PDFs
- editing DOCX files                    
- cleaning up Downloads folder (drag, rename, delete)

https://lnkd.in/ekzNWdXg

RICE and ICE frameworks optimize for the wrong thing.

They treat roadmap items as independent. Ship highest score, then next, then next.

But your roadmap is a system. Each item changes conditions for every other item.

Better question than "what's most valuable next?": What creates the most leverage for everything that follows?

- Best starting point is often NOT the highest-priority item - it's the foundational piece that makes everything else easier
- Some "low priority" items are root causes driving multiple high-priority outcomes
- Map which items have outgoing influence vs incoming effects
- Fix one driver and three problems disappear

Same items, different sequence. The team sequencing for reinforcement ships faster because each item makes the next one easier.

I codified this sequencing methodology as an AI skill for Claude Code or a prompt for ChatGPT. Transforms unstructured task lists into momentum-generating sequences using interrelationship diagrams.

Open article: 180+ AI skills that turn real product work into repeatable workflows by prodmgmt.world

180+ AI skills that turn real product work into repeatable workflows

How many great product ideas have you had that died because you couldn't build them?                                         

• Claude Code (my preference, not sponsored) eliminates the implementation gap                                                                              
• You get an idea → you can prototype it that afternoon                                                                      
• No more "I wish we had a tool that..." - just build it                                                                     
• No more waiting on eng bandwidth for simple automations                                                                    
• Start here: prodmgmt.world/claude-code                                                                                     

The best PMs I know ship things, so ship. It doesn't really matter what, don't overthink it.

While on a break I made:

- a BookTok to Goodreads PWA that helps me add books from images to my queue
- updated my personal site nurijanian. com with a lot of new content and most importantly 4 different styles that randomly rotate
- made lockinfocus. app for focusing before a deep work session using Andrew Huberman's insights
- made a better tab manager for Chrome that helps me sort my tabs by domain and even chat to the tabs with AI and sort them based on my stated goals
- gamified my goals with a personal dashboard that uses Supabase as backend

these apps will never be monetized or even public, but I made them just because I wanted to

now I am working on 2 new actual products:

- a car rental vertical SaaS that capitalises on my yield optimisation background
- an onboarding software that explores what an AI-first/mobile+web synced onboarding tool would look like

the sky is the limit

You're the most productive PM I ever met... and you're too busy copy-pasting between ChatGPT sessions to notice you rebuilt the same context many times over                    
                                                                                                                                                                                          
Most PM work is this loop: refine a process, explain it to ChatGPT, explain it again to Claude, explain it to your team, explain it again because you forgot the details. Same explanation, reconstructed from memory every time.                                                                                                                                      
                                                                                                                                                                                          
The issue isn't individual skills or Custom GPTs. Those are single-shot tools. The unlock is chained workflows where each step feeds the next automatically.                            
                                                                                                                                                                                          
Example ↓

everyone working with AI coding tools should watch this (41 mins)                                                                                                                       
                                                                                                                                                                                          
Geoffrey Huntley and Dexter Horthy break down the real architecture behind Ralph Wiggum loops - why the official plugin isn't it, and how to actually build AI coding workflows that work
                                                                                                              
I distilled the key techniques, but you need to watch them debug this live:

https://lnkd.in/eU3zeFH4

how AI has changed PM work, from an old-school PM on Reddit

less labor now, more speed

glad pre-AI PMs are embracing this so well, not enough cases like this

prodmgmt. world now supports .md files you can use with AI agents

grab the prompts, load them into Claude Code as skills
now you have an army of product managers doing your bidding

some of my favorite use cases:

- impact sizing the specs you're developing with AI
- researching users (assumptions, interviewing, analysis)
- product thinking (via Negativa aka inversion is my favorite)

collecting my favorite resources that I personally use with Claude Code as a PM

+ the guides I've written and the people who post about Claude Code on X

Bookmark it -- updating it daily!

🤾 prodmgmt.world/claude-code

taplio