Team Internet

Product Strategy & Vision
•	Developed 3-year product strategy, aligning portfolio initiatives with business objectives.
•	Led discovery processes to identify market opportunities, resulting in 32+ product launches.
•	Established product management way of working, improving prioritisation effectiveness by 40%.
•	Created standardised product management practices, enhancing team efficiency and collaboration by 50%.
Product Development & Delivery
•	Managed end-to-end product lifecycle.
•	Established discovery and delivery rhythms, reducing time-to-market by 35%.
•	Developed and maintained product roadmap.
•	Created and validated product requirements through user research and analysis.
•	Introduced new design language and processes, achieving 100% improvement in design consistency.
•	Created all new designs for each initiative. 
Customer & Business Impact
•	Conducted user research and synthesized insights to inform product decisions.
•	Optimised customer journeys, resulting in improvement user experience.
•	Created and tracked success metrics across product portfolio.
•	Analysed product performance data to identify optimization opportunities.
Stakeholder Management & Leadership
•	Led cross-functional teams across engineering, and business units.
•	Established strong relationships with key stakeholders through regular communication and alignment sessions.
•	Introduced Agile methodologies resulting in 30% increase in team output.
•	Created and maintained product documentation and communication frameworks.

Senior Product Manager (Australia)

Swyftx

•	Established product management best practices, improving team efficiency and collaboration by 20%.
•	Created and maintained product documentation and communication frameworks.
•	Established strong relationships with key stakeholders through regular communication and alignment sessions.
•	Delivered integration with NZ Merchant Payment Provider, enabling international expansion.
•	Reduced technical debt by 20%, enhancing system performance and reliability.
•	Mentored junior product managers and provided guidance on best practices.

Product Manager

Genie Solutions

•	Delivered a $5M product upgrade across three teams, aligning with strategic business goals.
•	Introduced product management processes, leading to a 20% increase in team productivity.
•	Managed multiple roadmaps and backlogs, achieving improvements in task prioritisation.
•	Streamlined release activities, reducing production issues by 30%.
•	Facilitated Scrum ceremonies, boosting cross-team collaboration and transparency by 25%.

Patient Zero Pty Ltd

•	Conducted requirements analysis, increasing accurate translations and product success by 30%.
•	Facilitated Agile ceremonies, driving 35% improvement in collaboration and client satisfaction.
•	Delivered key projects, including a fitness class scheduling rebuild for a $1B industry leader, resulting in continued client partnerships.
•	Optimised a sales app for payment and conversion rate improvements, enhancing customer journey experiences.

Product Owner

Mining Leaders Group

•	Designed and launched the Mining Leaders Accelerator, leading to $500K in equity acquisition and a 25% increase in startup success rates.
•	Built an executive leadership self-assessment tool, expediting decision-making for leadership teams.
•	Implemented digital marketing campaigns, increasing effectiveness and engagement by 40%.
•	Revamped company processes, improving operational efficiency.

Business Analyst (Agile) (Contract)

Merakium

- Managed international cross-cultural team of 8, overseeing content creation, development, and marketing.
- Developed strategic, marketing, business, and operational plans, improving organizational alignment.
- Designed workflow and process mapping, enhancing process efficiency by 30%.

Manager

Bond University

Program Manager

ilab Accelerator

Program Coordinator

I help SaaS founders and product teams turn feedback into amazing products

Tres West

Mission → 1k teams to adopt weekly prod. discovery. 📈 curr: 15/1000

I use AI tools 40h/week for real work. I share what actually works.

Most AI content is either too basic ("copy this prompt!") or too technical (47-node agent workflows). I'm in the middle — practical workflows for smart people who don't want to code.

**I post about:**
• Claude Code for non-developers
• MCP integrations that extend what AI can do
• Productivity workflows I actually use

**What I'm building:**
juttu — helping SaaS companies grow revenue from existing customers.

SWE → PM → Founder. Munich.

Niko Noll

I moved my company out of Notion.

Now my docs update themselves.

Every startup I've worked at had the same problem. You write docs during onboarding, during planning, during a sprint. Then they sit there. Nobody updates them. Six months later they're wrong and everyone knows it.

Notion made this worse, not better. Beautiful docs that are just as dead.

Last week I moved everything to Obsidian. Not because the editor is better. Because the files live on disk.

Eli Mernit put it well. The architecture of an AI agent is just two things: the filesystem as state, and Claude as the orchestrator. If your docs are files, agents can work in them.

So I added agent.md files with prompts that tell Opencode/Claude Code what each doc is for and how to maintain it. The agents clean up outdated sections. Flag inconsistencies. Update references when something changes.

The docs stay alive because somthing is actually maintaining them.

We sync them over Obsidian sync.

Most documentation problems aren't writing problems. They're maintenance problems. And maintenance is exactly what agents are good at.

What's the last internal doc you actually trusted?

P.S.
Hey, I'm Niko.
I'm building Product Analyst AI and sharing how I use AI workflows and product analytics to build better products.
I also make walkthroughs and tutorials on YouTube.
Follow along if that's your jam.

I share how I use AI to build, measure, and learn faster | Founder, Product Analyst AI

Famous VC built his own CRM with AI.

He'll be back on HubSpot next month.

Michael Bloch spent 50 hours building a command center with Claude Code. Email, calendar, CRM, the whole thing.

Uses it twice a day. Saves 15 minutes.

The build was easy. Maintenance sucked. Every new feature made the system more fragile. No one maintaining it but him and a chatbot.

Sounds familiar?
Yes, I'm looking at your 34 half backed lovable projects.

Everyone's asking "will SaaS die?" Wrong question.

The real shift is supply explosion. When building costs approach zero, every vertical that had 4 competitors gets 40. Every niche that wasn't worth a venture-backed company is now viable for a lean team with AI tools.

I'm one of those new entrants. No VC funding. Small team. Building Product Analyst AI with deep domain understanding of how tech businesses use data to make more money from existing customers.

For the consumer, this is great. More specific solutions that actually solve your problem.

For the builder, the job is the same as it ever was. Solve a specific problem deeply. But the bar is higher now.

We like it that way.

Don't just ship software that theoretically helps someone get the job done. Actually get it done for them.

Code got cheap. Solving real problems didn't.

What's the one thing your current tools still don't actually solve for you?

P.S.
Hey, I'm Niko.
I'm building Product Analyst AI and sharing how I use AI workflows and product analytics to build better products.
I also make walkthroughs and tutorials on YouTube.
Follow along if that's your jam.

Anthropic just launched a legal product.

Every B2B SaaS should be paying attention.

Not because legal tech is your space. Because a model company just decided the application layer is theirs too.

Think about what that means.

The company making the AI decided it can also make the product. No wrapper needed. No middleman. They went straight to the end user.

Will they do this for analytics? For project management? For CRM? Nobody knows. But the playbook exists now.

Meanwhile, MCP is making data portable. Users can query your product from a terminal. "Death of SaaS" is trending every other week.

Some of this is overhype. Some of it is very real. And right now, honestly, I can't always tell which is which.

We're building Product Analyst AI in the middle of all this. Some mornings it feels like we're building on quicksand. Other mornings it feels like the biggest opportunity of our careers.

Both feelings are correct.

The founders who win this moment won't be the ones who predicted everything right. They'll be the ones who kept building while staying honest about what's changing.

That energy is something else right now.

Are you more excited or more terrified? Genuinely curious.

P.S.
Hey, I'm Niko.
I'm building Product Analyst AI and sharing how I use AI workflows and product analytics to build better products.
I also make walkthroughs and tutorials on YouTube.
Follow along if that's your jam.

Last week I killed our Notion.

Here's what actually replaced it.

It starts with one file. We call it AGENT md. It's the map of our entire company. Folder structure, naming rules, key documents, and instructions for how agents should work with all of it.

The vault is one folder per area. Growth, Product, Customers, Weekly planning. Each folder has living docs that get updated continuosly and task files that get archived when done.

Three rules keep the whole thing alive:

1. Update before create. Search for existing docs first. If it exists, update it.
2. Archive don't delete. Completed work moves to _archive. Nothing disappears.
3. Link to parent. Every doc connects to its strategy doc. More connections = better context for agents.

Start of week: update This Week md with priorities. During work: create task files, link to parent docs. After completing: update the living doc, archive the task.

AI can generate endless documents. The hard part is keeping them current so the next session builds on the last one.

Full AGENT md in the comments if you want to steal it.

Are you also struggling with so much LLM output going nowhere?
This doc structure helps us, but always happy to hear suggestions on what to imrpove.

P.S.
Hey, I'm Niko.
I'm building Product Analyst AI and sharing how I use AI workflows and product analytics to build better products.
I also make walkthroughs and tutorials on YouTube.
Follow along if that's your jam.

We stopped asking AI to query data.

We taught it to think like an analyst.

First version of Product Analyst AI was simple. Connect to your data, ask questions, get answers. SQL under the hood.

It worked for "how many users signed up last week." Fell apart for anything real.

"Which users are about to churn?" It would write a query. Count logins. Return a list. Technically correct. Completly useless.

A real product analyst doesn't start with the query. They start with context. They know your product's lifecycle. They know what an AHA moment looks like for your users. They know which behaviors actually matter vs which ones just look busy.

So we built that into the agent. A context layer.

It knows your ontology. Your activation signals. Your key-value moments. The stages users move through before they upgrade or leave.

Now when you ask "who's about to churn," it doesn't just count logins. It looks at the patterns that actually predict churn for your specific product.

Same AI model underneath. Completely different answers.

If you're building AI products, this is the real unlock. Domain context beats raw intelligence. Every time.
And if you can combine the two, sure even better!

What domain knowledge would make your AI tools actually useful?

P.S.
Hey, I'm Niko.
I'm building Product Analyst AI and sharing how I use AI workflows and product analytics to build better products.
I also make walkthroughs and tutorials on YouTube.
Follow along if that's your jam.

Karpathy called it a phase shift.

I'm old enough to know what he means.

Andrej Karpathy, ex-Tesla AI head and OpenAI founding team, wrote that he went from 80% manual coding to 80% agent coding in weeks. Called it the biggest change to his coding workflow in two decades.

I'm not ex-Tesla, nor part of OpenAI. But I relate to all of it.

I build Product Analyst AI mostly in English now. 
I describe architecture, a system.
So do coding skills atrophy? Is my ego as a Software Engineer hit?

It's true:
My fingers genuinly hesitate on syntax I used to type without thinking.

But here's the part I'd add.

The engineering job didn't shrink. It shifted up.

I'm more of an architect now than a coder. I design systems. I make trade-off decisions. I think about how components fit together. That was always part of the job as an engineer. Now it's the main part.

I don't write individual functions by hand for well-proven, well-documented patterns anymore. That's the real change.

It's not slop. It's a higher level of abstraction.

And the expansion is wild. I build things now I never would have attempted a year ago. Not just faster. Things that weren't worth trying before.

Is your role shifting up or does it feel like something else?

P.S.
Hey, I'm Niko.
I'm building Product Analyst AI and sharing how I use AI workflows and product analytics to build better products.
I also make walkthroughs and tutorials on YouTube.
Follow along if that's your jam.

Half of AI agents just write code.

Every other work is still being done manually.

Anthropic published their agent deployment data. Software engineering takes roughly half of all agent usage.

Marketing, ops, data analysis, sales? All under 10%.

Greg Isenberg on this chart: "every category under 10% is a department still running on manual effort and tribal knowledge. Turn one of those workflows into a repeatable AI agent. That's vertical SaaS 2.0."

Whatever you think about AI hype. Agents will be built for all of these segments. The opportunity is figuring out how.

This is what excites us at Product Analyst AI. How do you build an agent that acutally gets the job done in analytics? Not just querying data. Helping you find signals, improve your product, make better decisions.

Every product person should be asking this right now. What does a useful agent look like for your vertical? What work does it do?

Most of these segments are wide open.

What vertical would you build an agent for first?

P.S.
Hey, I'm Niko.
I'm building Product Analyst AI and sharing how I use AI workflows and product analytics to build better products.
I also make walkthroughs and tutorials on YouTube.
Follow along if that's your jam.

Every startup we worked with think they have a data problem. They don't. 

They have a setup problem. The data is there, it's just not structured for the questions they're actually asking.

And AI makes this harder to see, not easier:
You type a question, get a confident answer, and assume the analysis is sound. But the quality of what comes back depends entirely on what's underneath.

There are 3 levels of data questions. 
Each one needs a different setup to actually work.

Level 1: Query.
Count, filter, group. One question, one answer.
"How many trial users converted last month?" → 1,247.

Level 2: Analysis.
Compare, correlate, find patterns across behaviors.
"What do users who convert have in common that those who don't, don't?"

Level 3: Prediction.
Score users based on behavioral patterns over time.
"Which free users are most likely to upgrade this month?"

Most teams are at Level 1 and think they're at Level 2. Because the interface feels smart even when the operation underneath is simple.

The AI isn't the bottleneck. Your setup is.

We're building the layer that gets teams from Level 1 to Level 3. That's where the real insights live.

Which level are most of your data questions at?

P.S.
Hey, I'm Niko.
I'm building Product Analyst AI and sharing how I use AI workflows and product analytics to build better products.
I also make walkthroughs and tutorials on YouTube.
Follow along if that's your jam.

"AI > UI" was a spicy take in 2024. In 2026 it's just... what's happening.

If your product's value lives in the interface, that's worth sitting with for a minute.

Amplitude now has an MCP server. Connect it to Claude Code and just ask questions. "Which free users are closest to converting?" No dashboard. No filters. No segments. Just the answer.

Incredible for users. Less so if your product IS the dashboard.

The tension is real.

Open your data via MCP and users query it from any AI interface. Your UI becomes optional. Your edge shrinks.

Keep it closed and you're the walled garden nobody wants.

Amplitude chose to open up. Smart move. But the value now has to shift. From presentation to curation. From "here's a dashboard" to "here's what you'd never think to look for."

Dashboards won't disappear. You'll still check the same five charts every Monday.

But the ad-hoc question, the "I wonder if..." moment... these are moving to chat. And they're not coming back.

The products that thrive won't fight AI for the interface. They'll be the data layer AI can't work without.

Would you give users MCP access to your product's data, or keep the UI edge?

P.S.
Hey, I'm Niko.
I'm building Product Analyst AI and sharing how I use AI workflows and product analytics to build better products.
I also make walkthroughs and tutorials on YouTube.
Follow along if that's your jam.

Build-measure-learn is about to close itself.

OpenAI just showed us how it starts.

A small team shipped an internal product. Zero manually-written code. A million lines. ~1,500 pull requests. Three engineers. 1/10th the time.

But the code generation isn't the story.

The agents closed their own loop. Boot the app, drive the UI, reproduce a bug, record a video of the failure, fix it, open a PR, merge. While the humans slept.

Write. Test. Observe. Fix. No human in the middle.

Right now that loop is narrow. "Measure" means test results and CI passes. "Learn" means fixing what broke.

But imagine widening it.

What if agents could see usage data? Not just "does the code compile" but "are users actually clicking this button?" What if experiement results and funnel drop-offs fed directly into the agent's context?

That's build-measure-learn fully closed. We're not there yet, honestly. But the direction is clear.

This is a scary and exciting time to build. I'm choosing exciting.

What's the first piece of your product workflow you'd want an agent to close the loop on?

P.S.
Hey, I'm Niko.
I'm building Product Analyst AI and sharing how I use AI workflows and product analytics to build better products.
I also make walkthroughs and tutorials on YouTube.
Follow along if that's your jam.

taplio

The Battle

Tres West

Niko Noll

Followers

Posts

Engagements

Likes

Performance ScoreNiko Noll's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Tres West

Niko Noll

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Niko Noll's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreNiko Noll's LinkedIn Performance