Dr Bart Jaworski

You might have felt that Jira is the source of product management issues. Now it can become your product management central hub. Here's how: At Skype, we got feedback coming in with what felt like trucks. The best I could do was randomly look at it, hoping to see some patterns and inspiration. However, the problem was clear: there was too much feedback to make it actionable. Here's what most product teams know in their bones. The backlog isn't where product work starts. It's where it ends up. Every ticket in Jira traces back to inputs that live somewhere else: app store reviews, your feedback widget, research sessions, the help desk, stakeholder asks. The signal is scattered across a dozen places, and turning it into a roadmap usually means dragging data from tool to tool until it finally lands as a sprint. For years, product managers borrowed engineering's tools and made them work. We never really had a stack built for us. That's what's changed. Atlassian's Product Collection is a set of connected tools designed around how product teams actually work, distinct from Jira rather than bolted onto it: 1) Feedback captures and organizes customer signals from all those scattered sources, then uses AI to surface searchable, actionable insights. The mountain I drowned in at Skype? This is the tool I needed back then. 2) Jira Product Discovery turns insights into roadmaps, with built-in frameworks to prioritize ideas. It helps you align stakeholders around the right decision, so you're not rebuilding five versions of the same roadmap when goals shift. 3) Rovo works alongside you, surfacing insights, drafting PRDs, and connecting strategy to delivery, so engineering understands not just what to build, but why. 4) Finally, with the Pendo integration, teams can connect what customers are saying with what they are actually doing in the product. That's the full arc of product work, from raw signal to shipped decision, in one place. Honestly, the tool I wish I'd had back at Skype. See what Atlassian built for product teams: https://lnkd.in/dq74xqrH So here's my real question: where does your team's feedback actually live right now? Let me know in the comments. #productmanagement #AtlassianPartner #𝗔𝗱

You probably felt mentally checked out when management asked you to work on a project you know is going to fail. Here's what you do: Well, in short, you make the best of it. I presume that you did your best to say no, to show metrics that prove that there are better things to do, that you've been in a big fight over your backlog, and still you were "asked" to work on the ridiculous request anyway. Here's how to proceed in this uncomfortable situation: 1) Keep your cool Of course, it all depends on whether this is a one-time thing or you have no space to work on your own ideas and reach your goals as you see fit. Sometimes bad ideas are simply forced upon you. I know, I know, it goes against every product management principle I'm showing here on LinkedIn. However, since those things sometimes happen, the best you can do is not worry about it and just do your job. And by that I mean: 2) Get the most value out of a bad idea. There are two ways of looking at this: You either turn a bad idea good, or at least a promising one, or you put the least possible resources on it and create an MVP that will fail so quickly and so spectacularly that you can move on to your regularly scheduled roadmap. Basically, rather than despair, turn it into a challenge. 3) Don't show your discontent. You can't just express how you hate the project to your team and to everyone else. For one thing, it's unprofessional. And if your team sniffs out your bad attitude, they will definitely copy it. Instead, it's better to try to force a false optimistic narrative in order to, perhaps, hear an idea that will flip that bad proposal into a good one. Who knows? Perhaps this AI project that would consume millions of dollars in tokens can be optimized to be a neat feature that consumes very few tokens, and one of your teammates knows how to do it? 4) Just in case... keep the paper trail. In case you find yourself in the middle of a political power struggle where a bad project is being used to undermine someone, or even you, it's best to be able to prove that this bad project was not your idea and initiative in the first place. You don't want something you are not in control of to come back and bite you later. 5) Once it fails, record the learnings. So, once the poo hits the fan, all you can do is to make sure that no more poos are thrown in the fan as much as possible. Just document the failure, explain what happened, why it happened, and move on to something better, so you can better showcase your product management talents in your company. Hope this helps next time you are put into this situation. What was the worst update you had to develop that was forced on you? Let me know in the comments.

Are you confused about how to become an AI PM? Here's where to start, even without a coding or technical background: 1) Master the right skills. Technical: • LLMs, tokens, and context windows • RAG, embeddings, and retrieval quality • AI agents and tool calling • Evals and LLM-as-judge • Prompt architecture • Latency vs. quality tradeoffs • Inference costs • Hallucinations and guardrails • Fine-tuning vs. prompting vs. RAG • Human-in-the-loop workflows Soft: • Executive communication • Systems thinking • Strategic framing • Cross-functional influence • Risk communication • User empathy • Decision-making under uncertainty • Storytelling • Stakeholder alignment 2) Build proof-of-work projects. You learn by doing. Period. Build things that prove you are AI PM through and through. Examples: • meeting notes to action-items converter • churn-risk signal detector from support logs • automated competitor changelog tracker • RFP/security questionnaire drafting assistant    With that, build a portfolio, where you document the problem, the workflow, the prototype, the model choice, the eval plan, the cost tradeoffs, the failure modes, the UX decisions, etc. next up: 3) Fix your CV and outreach. Your resume should not say: “Used AI to become AI PM.” Dull. It should say: • Shipped an AI triage system that cut first-response time on support tickets by 45% • Built an LLM-as-judge eval pipeline to catch hallucinations before release • Automated weekly exec reporting, saving the team ~8 hours a week Your CV needs to scream, “I can create AI delivering real value.” And for outreach? Well, you already built your portfolio, didn't you? Send hiring managers: • a sharp AI product teardown • a prototype relevant to their company • a 1-page AI product strategy • a short Loom demo 4) Push further: Create your own AI opportunity Find a problem worth solving, build a prototype, and show how it provides value. Do it for your current company (that can promote you) or when applying for a new job (and build the prototype for that company's product). You can then say: “I think this could become an AI-native workflow for our team. I’d love to own the next version.” That is how internal transitions happen. >>> If you want to learn all of this without wasting time jumping between hundreds of YouTube videos, outdated resources, and advice from AI influencers who aren’t actually building, you’ll love Product Faculty’s #1 AI PM Certification. You’ll learn directly from Codex’s PM and other frontier AI operators who are building real AI products. Inside the certification, you get: • Live sessions • AI Build Labs • Build your capstone project with 1-1 support • Exclusive AI PM content library • Enterprise AI systems • 1:1 support (3,000+ students, 1,000+ reviews.) Enroll here for $500 off: https://lnkd.in/ds59DbeQ P.S.: Comment "AI PM" and I'll send you free resources.

Did a translation ever delay an important release for you? Well, you no longer need to wager quality vs timely release! Here's how: As a person who used to work on Skype and shipped the product for tens of languages, I personally know how difficult it can be sometimes to spend months on a piece of work and then have it delayed because the translations are missing. And very often there aren't too many strings. You can just pop in what you need to translate there and ship it off. But do you know what you're shipping? Is it good enough? Aren't you offending anyone by accident? I have a solution for you. If you ship in more than one language, QACAT from my patron, Alconost, is live on Product Hunt today, check it out: https://lnkd.in/dXJqBPQw For years, localization QA meant one thing: spreadsheets. Marked-up strings, a column of error notes, a final score nobody fully trusted. Slow, awkward, and the first thing to get cut when the ship date is breathing down your neck. Then AI made translation fast and cheap. The catch? It also made quality invisible. You get output in seconds, but no real read on whether it's actually right. So teams keep hitting the same bad choice every release: ship on time and hope the translations hold up, or hold the release for a proper review. QACAT, launching today on Product Hunt, takes that choice off the table. It's a hybrid translation QA platform from Alconost that lets you match the QA depth to the moment: • Rule-based checks when you need speed (fast, and NDA-safe). • AI analysis, or AI plus human review, when you want more confidence. • Full expert human evaluation when the stakes are high. A few things that stand out if you've ever lived this process: Review happens on real screenshots, not exported strings. Built-in OCR pulls the text for you, the translation auto-fills, and reviewers mark issues right on the UI. Every run produces a structured, scored report: severity breakdowns, error categories, language and engine splits, plus an AI summary that points straight at what needs fixing. And quality becomes trackable over time, so the same issues stop resurfacing release after release. 100+ languages, one environment. That's the whole point: speed and quality stop being a tradeoff. You dial QA up or down to fit what each release actually needs, instead of betting one against the other. What's the funniest bad translation you have ever seen? Let me know in the comments.

🍎 Apple is hiring an AI PM. Even Netflix is paying $900k/yr for the AI PM roles. Here are the good resources to become an AI PM: let's dive into the resources: 𝟏. 𝐁𝐚𝐬𝐢𝐜 𝐂𝐨𝐧𝐜𝐞𝐩𝐭𝐬 Start with what an AI PM is: https://bit.ly/whatisaipm Next, for most PMs, it makes no sense to dive deep into statistics, Python, or loss functions. Instead, read about Transformers, and LLMs: https://bit.ly/3EZtCLs 𝟐. 𝐏𝐫𝐨𝐦𝐩𝐭 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫𝐢𝐧𝐠 Free resources: 🔗 GPT-4.1 Prompting Guide 🔗 Anthropic Prompt Engineering 🔗 Prompt Engineering by Google 🔗 System Prompt Analysis for Claude 4 🔗 Anthropic Prompt Generator 🔗 Anthropic Prompt Library 🔗 Prompt Engineering Course By Anthropic All links: https://bit.ly/pcprompts 𝟑. 𝐅𝐢𝐧𝐞-𝐓𝐮𝐧𝐢𝐧𝐠 Learn by doing. No coding: 🔗 OpenAI Platform (start here) 🔗 Hugging Face AutoTrain (best for other models) 🔗 LaMA-Factory (fine-tune open-source LLMs) All links: https://bit.ly/pcfinetune 𝟒. 𝐑𝐀𝐆 (𝐑𝐞𝐭𝐫𝐢𝐞𝐯𝐚𝐥-𝐀𝐮𝐠𝐦𝐞𝐧𝐭𝐞𝐝 𝐆𝐞𝐧𝐞𝐫𝐚𝐭𝐢𝐨𝐧) RAG, by definition, requires a data source + LLM. But there are dozens of possible architectures. 🔗 A free, interactive RAG simulator: https://lnkd.in/dbERNB8u I also recommend a simple step-by-step exercise to build a RAG chatbot in practice. No coding: https://lnkd.in/dZ-e_C9G 𝟓. 𝐀𝐈 𝐀𝐠𝐞𝐧𝐭𝐬 & 𝐀𝐠𝐞𝐧𝐭𝐢𝐜 𝐖𝐨𝐫𝐤𝐟𝐥𝐨𝐰𝐬 My favorite tool, by far, is n8n. You can host a free version locally or in the cloud. Start with those guides: 🔗 MCP for PMs 🔗 Automate Anything with n8n 🔗 AI Agent Architectures My favorite free resources: 🔗 Google Agent Companion 🔗 Anthropic Building Effective Agents 🔗 IBM Agentic Process Automation All links: https://bit.ly/pcaiagents 𝟔. 𝐀𝐈 𝐏𝐫𝐨𝐭𝐨𝐭𝐲𝐩𝐢𝐧𝐠 & 𝐀𝐈 𝐁𝐮𝐢𝐥𝐝𝐢𝐧𝐠 My default no-code tech stack: Lovable, Supabase, GitHub, Netlify, n8n, Stripe. Four practical tutorials: 🔗 AI Prototyping 🔗 How to Quickly Build SaaS Products With AI 🔗 How to Build a Full-Stack App with Lovable 🔗 No-Code B2C SaaS Template With Stripe Payments All links: https://lnkd.in/dt_q7RQC Thank you, Paweł Huryn for curating and writing these resources! Do you want to have a great strategy to land your next product job? Get a new PM job in 2026 with Aakash Gupta and my cohort "Land PM job". We're starting the 4th cohort very soon, after already getting many people hired from the previous ones. Check it out here: www.landpmjob.com #productmanagement #aipm ##ai

Context switching can be a real pain for a product manager who needs to use 20 different tools to do their job. Here is how to bring them all to a single, efficient place: I have been a big fan of Miro ever since they were just beginning. It was a pleasure to use the software, and it helped me massively in many brainstorming sessions. However, there was always this issue of switching between Miro, Jira, and whatever other source of context, data, and inspiration I would need in order to effectively collaborate with my team. We've been watching Miro's AI evolution over the years, and the way they could already pull context from different sources was impressive from the start. And they are not backing down. I was invited to the keynote presentation in London for Canvas 26, and what they've shown as the next evolution of Miro is yet again an excellent step in creating a Product-Management one-stop suite. Let me focus on one of the items of the presentation, which impressed me the most: Making Miro the main, perhaps single, source of truth. Their AI assistant, called Sidekicks, will soon be able to pull what's relevant from Jira, Confluence, GitHub, Slack, and more, and act on it without you having to manually move data between tools. Can you imagine turning a Slack discussion into a usable brainstorming workspace? It's now possible and quite seamless, really. All you need to do is let your Sidekick know that this discussion took place, and it will do the work for you. You don't even need to show AI where the discussion happened. It can find it on Slack on its own, and reliably, might I add. When you're done and you know what you want to build, you don't need to switch back to Jira and create a ticket. The Sidekick will do that for you when prompted. All you need to do is be the best product manager you can be, lead with your creativity and experience, and then just make sure that what AI created is what you wanted to see there. The more tools you connect, the more your Sidekick can do: from answering questions with live data to taking action across systems on your behalf. And that's just a single update out of many shown at Canvas 26. If you want to learn more and see how easy and efficient it is to be a product manager with Miro, check out the link below: https://miro.pxf.io/m4EL37 What's the one tool you wish would just talk to all your others? Let me know in the comments. #miropartner #productmanagement #productmanager

Are product managers on the brink of becoming one-person product creators? Or did we just hit a sobering reality check? Here's how it is: This isn't another "AI will replace you" post. It's about where I think things actually stand, and what might come next. AI was sold as an almost omnipotent stand-in that could replace anyone, fast and cheap. The pitch looked plausible at first. In reality, the cracks showed quickly. Models have improved a lot since ChatGPT launched, but nowhere near the point of replacing people. Some limitations that make this impossible are baked into the design. Hallucinations won't ever fully go away. Neither does the tendency for an LLM agree with you. Klarna is the cautionary tale. It replaced around 700 customer service staff with AI, then started rehiring within roughly a year. The CEO admitted the chatbots were cheaper but delivered lower-quality service. MIT's "GenAI Divide" report found that ~95% of corporate generative-AI pilots deliver no measurable impact on the bottom line (a figure that's since been contested, though the direction holds). The 5% that work understood the assignment: companies like Slack and Miro build AI that enhances the work instead of trying to replace the worker. Then there's cost. Anthropic genuinely cracked usable coding agents, but at a price. Uber reportedly burned through its entire 2026 AI coding budget in four months. Microsoft pulled Claude Code from one of its divisions after token billing blew past the annual budget, moving engineers to its own tool. I tried a small side project myself and felt like I was paying every time I had a productive thought. It starts to look like manufacturing: some shops run expensive robots, plenty still rely on cheap manual labor. So, where does this leave PMs? For now, I see AI falling into three buckets: 1) Productivity enhancers. Note-takers, dictation, and whiteboards that pull project data from across the company. They quietly make us better at the job. 2) Faster time to market tools. Software like Lovable and Bolt compress prototyping, so we can test more directions before committing to one. 3) Strain amplifiers. Yes, really. If a team ships more code with agents than with people, PMs face more options to build than we can possibly validate. When you can quickly build any feature, picking the right one gets harder, not easier. You can't release 20 features a day. Users are still flesh and blood. They need time to adapt, learn, and onboard. After all, most of us build stuff for people, not agents. Not to mention that more features don't have to mean a better product. Bottom line: we're still in the "let's see what happens" phase. We need to observe carefully and ignore the wishful marketing. Taking AI companies at their word is a reliable way to get surprised by reality. What about you? Did AI impact the way you work? Let me know in the comments #ai #productmanagement #productmanagers

It's okay to lose focus on a meeting until you miss that one important detail that is vital to your product. Here's how to prevent this: Listen, I perfectly understand the situation. You are in your fifth or sixth meeting of the day, struggling to stay awake, not to mention vigilant. Yet, all of those meetings held potential to be very important. Fortunately, there is Granola to the rescue! Granola is an AI notepad that easily turns my meeting transcript into searchable meeting notes with action points and key conclusions. This means that even if I slip, I always have ways to fall back and recall the whole meeting. I recall being in a meeting where I felt like I had no place being, as it was a technical discussion between two engineering teams on how to implement a refactor of a specific feature. I could barely stay awake. However, I literally woke up to hear an important detail that scratched my brain and forced me to ask a fundamental question. That question eventually led to shortening the project from roughly four months down to two weeks, because of a simple solution that my question uncovered. We would have wasted so much time if it weren’t for that key moment. However, you might not always have that lucky wake-up moment. Granola makes sure I don't need one. What is particularly useful for me is the search function. It's very easy to forget or miss an important detail that can make a fundamental difference later. With this available, I never miss an important piece of feedback, a key action item, or a critical decision. Moreover, everyone in the room can pay attention because no one needs to be hunched over a keyboard and take notes, which means you get the most value out of each meeting. This especially helps my creator journey, as very often tiny contract details with clients are agreed on the call and get forgotten when it comes to signing a contract. With searchable meeting notes, it’s easy to show the partner the most important, missed details, rather than a game of “he-said, she-said”. Especially, that Granola works directly off your audio input, so no bots are required, and it will naturally fit into your process and day. Check out granola free for 30 days using my link: https://lnkd.in/d8pvUYdi What was your meeting where one important detail turned things on its head? Let me know in the comments. #productmanagement #productmanager #sponsored

Do you also struggle to translate data into understandable insights, not to mention a story? Here are 7 ways to achieve this: 1) Focus on what data means, not what it is I will be the first to admit, it is easier said than done. Oftentimes, both will be the exact same thing, but remember, it's about how you approach it. You need to be a storyteller, not an accountant. Something along the line: "20% more people enjoy our shows when compared year to year". vs. "Our MAU is up 20% YoY". Of course, the first one will sound great on a live presentation, and the 2nd one will be a better choice for a concise update email report. However even if the right language is chosen, you are still in a risk of burying the recipients under a ton of updates, thus: 2) Follow the rule of three Whether it is a slide or an email, try to focus on only 3 key figures or at least 3 metrics per metric category. Maybe you can take it up a notch and even use a single, most important indicator? Even then, help yourself out and: 3) Use the fonts and colors to address the importance and impact That's right! If something is huge and growing, make it HUGE and green! 3) One chart at a time! While you may think a lot of charts is cool, they can be confusing. Instead, have a single graph at a time, and if you need to show information visible only between different charts, overlay it slide by slide. 4) Ensure the technical data foundations first No one wants to report something that turns out wrong. Know your certainty level, standard deviation, and make sure that your calculations are stellar. You never want to be in a situation where you have to walk back good news (I did...) 5) Skip the details Just focus on the jist of it. But, by all means, have the details ready to share upon request or via emails/articles. Just don't pre-emptively answer questions that may never come! 6) Talk only about solid outcomes A metric improved 0.5%? Cool, cool, cool, cool... But there is a potential 2% frame of error, right? If this change is so small, it's unlikely it is a success worth bragging about. 7) Make it accessible Data should be available to anyone! Share your dashboards and make them easy to find. Maybe someone curious may notice something you missed. Do you usually struggle to tell a story with data? Sound off in the comments! #productmanager #productmanagement #data

How is it that, as a product manager, you have to abandon most of your work half-baked? How to escape the MVP trap? Stop me if this sounds familiar: You and your team spend months on an incredible idea. You launch the MVP. It shows great results. Of course, there is so much more you can do to improve it. And then it happens... You suddenly need to abandon your current success and work on another MVP to have a shot at hitting, representing mounting stakeholder pressure. The cycle repeats. Result? • Tech debt and bugs are piling up • Half-baked features that never reach their potential • Disjointed products, because “we’ll integrate it later.” • Promising features were tossed aside when they only needed refinement • A frustrated team that feels like they’re building sandcastles, not products Well, if it is this bad, why does this keep happening, you may ask. Well: • Risky bets get prioritized over safe, compounding wins → hype always outshines steady growth. No idea why dev cycles seem to be run by adrenaline addicts. • Market conditions and management change faster than dev turnover → strategy gets rewritten before work is finished. • Everything takes too long → by the time you’re ready to release one thing, priorities have shifted. • The toxic belief that “𝘯𝘦𝘸 𝘪𝘴 𝘢𝘭𝘸𝘢𝘺𝘴 𝘣𝘦𝘵𝘵𝘦𝘳” So, what do you do? How do you fight back to deliver quality, optimal performance, and the best products? Here are some suggestions: 1) Design MVPs that fit your product’s quality bar While MVP is a version of your product that can prove your hypothesis with the least effort, perhaps you have put more time and effort after all, as if this were something close to a final release from day 1? 2) Put improvements on the roadmap When stakeholders see that iterations are planned, they’re less likely to cancel them for the next shiny gem. 3) Leave sprint space to revisit older work Most teams fail here. By deliberately carving out capacity, you give features room to grow, polish, and integrate. Otherwise, every MVP is destined to rot in production and add to the graveyard of “good but forgotten” ideas. 4) Use transparent, data-driven prioritization If your follow-up work stacks better against new, risky ideas, it's easier to defend staying on the same dev lane. 5) Educate stakeholders on the cost of distraction If your managers and stakeholders only see successful MVPs, why wouldn't they push for more? They need to know the cost of such actions. 6) Build your independence and authority This is a story for another post. So, have you ever had to leave a new, successful feature in the middle of the planned development cycle? Let me know in the comments. #productmanagement #productmanager #mvp