Product Agility

Alex Howard: From Founder to Product Leader: The Series A Blueprint & Why Outcomes Matter in the AI Era - Productized 2025 TalkInTen

• Ben Maynard, Barbara Fazeka, Alex Howard

Share episode

Send us a text

Productized Lisbon is world-class - inspiring sessions, brilliant community and hands-on conversations. We’re honoured to partner with Productized for the third year in a row.

Welcome to a special Talk in Ten from the Product Agility Podcast, recorded live in Lisbon at Productized. In this rapid-fire episode, Alex Howard walks us through his Series A Blueprint, the idea of a fractional (and fractionalised) CPO, and why founders need product leadership skills now more than ever - especially in the generative AI era.

Key topics discussed

  • The Series A Blueprint: scaling product and team from pre-seed to traction
  • Fractional CPO vs. a do-it-with-you model that empowers founders
  • The rise of the "product founder" era and product-led growth
  • Why CEOs must be outcome-led in a world shaped by generative AI
  • Feature commoditisation, AI tooling and the importance of curatorial ability


Guest bio:

Alex Howard is a product leader who helps founders and CEOs move from early-stage product work to Series A traction. He runs a cohort-based Series A Blueprint that mixes frameworks, recipes and mastermind sessions to scale product leadership without stepping into a full-time CPO role. Alex also runs remote workshops and coaching for founders aiming to build stronger product muscles. See Alex's Responsible AI Podcast - https://open.spotify.com/show/63PqRkJ9sk9RnZNjAuxDOf?si=ab826224c6cf4eb7


Want more? Alex is hosting a remote workshop on October 30th - https://www.regenerativeproducts.io/series-a-workshop?utm_campaign=as-npt116018533

Huge thanks to our sponsors, Bobcats Coding, for making this Lisbon series possible. Bobcats is a Budapest digital product studio specialising in AI engineering and end-to-end digital product development. Download their AI economics guidebook at bobcatscoding.com.

Host Bio

Ben is a seasoned expert in product agility coaching, unleashing the potential of people and products. With over a decade of experience, his focus now is product-led growth & agility in organisations of all sizes.

Stay up-to-date with us on our social media📱!

Ben Maynard

🔗 https://www.linkedin.com/in/benmaynard-sheev/

🐦 https://x.com/BenWMaynard

💻 https://sheev.co.uk/

Product Agility Podcast

🔗 https://www.linkedin.com/company/productagilitypod/

💻 https://productagilitypod.co.uk/

🖇️ https://linktr.ee/productagility


Listen & Share On Spotify & iTunes


Want to come on the podcast?

Want to be a guest or have a guest request? Let us know here https://bit.ly/49osN80

Before we launch in, let me just for those of you who aren't already using Prodpad, let me just explain what what prodpad is. So we are a complete product management platform. We a tool for product teams to have a single source of truth for all things product. So it enables product teams to run their product strategy, their product planning, to build and communicate their product roadmaps, to capture, prioritize all their product ideas. And then also it is a central place for gathering and routing or customer feedback so that you can link what your customers are saying to the product ideas that you are prioritizing and what you're communicating on your roadmap. And of course, which we'll see a lot of today, prodpad comes complete no matter what package or plan you pick. For Pad, it comes complete with copilot, which is the world's number one product management AI tool built specifically for the product management context. Primed with product management, real experience and best practice know how. But I won't bang on too much about Copilot because you're about to see some of it in action. So before we, before we, before I pass over to Janna just to set the stage for today's discussion. So AI and product management. There are two key facets really of AI and product management. One, how you as a product manager, as a product person can use AI to make your day to day work more efficient and effective. And then second, how you can integrate AI into your product to sort of gain competitive edge, solve bigger and better problems. Just for clarity, we are going to be diving into the first area today. So this is about how you can use AI in your work, in your role, to work better, to work faster, to work smarter. If you do want guidance and advice, examples relating to that second part of how to manage an AI product yourself, how to build an AI product, AI features then we have covered that, we've covered that in a previous webinar and we have this sort of definitive ebook here that has everything. So you can use that QR code to download it, but I'll also include a link in the follow up email that we send to you later in the week. But on to Janna now to look at offloading your PM busy work to AI. Excellent. Thanks so much for the warm intro and kicking us off here. So you know, I want to talk about the real, the reality of product management. Like not all the work we do is equally valuable. There's a lot of stuff that we do but you know, really it's at the Top of this pyramid where we create the most impact. You know, setting vision, prioritizing the roadmap, making the strategic decisions, those trade offs, this is where you really earn salary. In the middle you've got, you know, all the research, the analysis, the planning. This is the work that supports those big decisions. It's still important, but it's more about feeding insight than driving strategy. And down at the bottom, which somehow takes up most of our work, seems to be all the admin, you know, writing specs and cleaning up the backlog, triaging feedback. It's necessary, but it doesn't move the needle by itself. And the problem is too many PMs are stuck here at the bottom taking up all their time with admin and repetitive tasks. And this is where AI can change the equation. Right at the bottom, let AI take the weight. Right in the middle you can work together with AI to. It's kind of like a tag team. But at the top where this real value is created, this should stay all you the more you can shift your time upward and spend more time at the top of that pyramid rather than down at the bottom doing all the repetitive tasks, the more valuable you can become to your company. And that's exactly what we're going to see play out when we follow our use case today. Ella. So meet Ella. Ella is a senior product manager at a company called Apartment Tracker. You might know Apartment Tracker is one of the accounts the, the products that we use as a dummy product in our demo accounts here at Prodpad. So you're going to see some ex examples coming out from there and her day looks a lot like ours as a product person, right? A constant juggle of feedback and stakeholders and customer issues and strategy, all while putting out fires left and right. Now she knows her job is to spend more time at the top of that pyramid on vision and strategy, but in reality, most of her day gets swallowed up by the admin at the bottom. So I want to talk you through Ella's day and show you how she uses AI, in this case prodpad Copilot. She's going to offload a bunch of busy work and you're going to see how she can sh. Shift her time upwards. So we're going to follow through a bunch of different crunch points and see the exact prompts she uses, the output she gets and the choices she makes to keep the strategic human work for herself. So let's start with how her morning begins. So she gets into work, she opens her laptop, it's 8:45 and straight away there's a crisis, right? You all know this one. An exec has emailed asking for a quick roadmap Update before their 9:30 meeting. Absolute classic, right? So our first thought is, I don't have time to pull that all together from scratch. And so this is the exact sort of busy work that you know can sink your morning. It's sort of thing that you could lose an hour digging around through Roadmap and rewording updates, formatting them into something stakeholder friendly and still rushing it out the door, hoping that it makes sense. So instead of scrambling, Ella drops a prompt into copilot. Can you summarize this roadmap as it stands and draft a stakeholder update email in my tone of voice, in my usual tone. Now, the reason why she's using Copilot here is because Copilot is a tool that sits within your prod pad, which is a place that captures everything from your vision to your goals to the things that are in your backlog to what your customers have asked for. And so it knows what your roadmap is. It allows you to summarize what's already there and then output instantly really helpful things that are going to help you get away from some of this, this grunt work. You could do this in a generic AI tool as well, ChatGPT, for example. But the difference is that you'd have to then give all this context to this tool in order to tell it what the roadmap is so that it can summarize it for you. I'm going to give you some tips on how to handle this either direction, but we're going to focus on speeding this up by showing how you do it with Copilot in these examples that Ella's going through. So Ella's used Copilot here and it's come back with a now, next, later summary summary, which includes a polished draft email that reads something like that she could have written herself because it knows her tone. So if she had just copy pasted this and sent it, she'd risk missing some context. So you can't just copy paste everything from AI. There might be a customer win from yesterday's release that she wants to add. You know, this is the trap of bad AI use is if you, if you end up just plopping prompts in and then sending that stuff right out, you save time, but you lose credibility, right? What Ella does, because she's good at using AI, she reviews the drafts, she tweaks bits and pieces, she adds that customer highlight from yesterday. And she makes it her own. And it takes it from 45 minutes of writing down to 5 minutes of writing by using AI to speed up. So, you know, the, the key here is that AI can be used to draft your work and then you, as the product person can curate. That's how you get that really fast, accurate on message updates without drowning in details. So stakeholder updates in a rush, really perfect for that sort of thing. So she sends off that stakeholder email and checks the clock and there's five minutes to stand up and she realizes that she hasn't touched the backlog in days. A bunch of feedback has been piling up. She's got customer tickets, interview notes, sales requests. It's all sort of dumped into the system. And she knows that she needs to go into this meeting and she's going to be asked, what's at risk? What's changed? What are people talking about? And the thought is, how can I make sense of all of this in just five minutes? And normally she'd have to scroll through piles of raw feedback and hope she spots a theme, but this time she opens up prodpad Signals. It's prodped's way of cutting through the noise to show you the themes that matter. And it's an inbuilt tool that already has access to all the feedback, all that stuff put into one place. But it will, with one click, summarize everything. It'll pull out the signals from the noise. And she could do this with one click, coffee still in her hand, and it's going to highlight things that are, that are trending. So for her product, perhaps it's things around the, the onboarding improvements where a whole bunch of points of friction were logged. There's a key feature, landlord reports, and there's some, some failures coming out of the, the export function just recently, maybe it's pulled out information about, you know, there's a bunch of missed alerts and so that's the signal, not the noise. She's now able to see this stuff and bring it to her stand up. She's not walking in, guessing or bluffing or without any information at all. She's able to walk in with the three biggest risks already on the tip of her tongue. So don't let raw feedback bury you. Signals does the heavy lifting so that you can lead the conversation with confidence. Now, again, this can be done with other generic AI tools as well. We've got tips on how to set up prompts to do this, but the key is, is making sure that you've got all of it in one place. And what we've done with Signals is turned it into a single button that you press, so it just looks through everything that you've got in there and tells you what, what you need to know. So standup wraps up and Ella thinks that maybe things are under control. But of course, she's a product person and the next fire lands in her lap pretty quickly. Support team flags her, right? She, she gets a ping and apparently it's a high priority customer who's furious. That landlord's report thing that we're talking about, it's breaking their exports and it's hitting multiple key accounts. And her thought is, oh God, this is bad. Like, I don't have time to read through every ticket or piece of feedback about this feature, but I need to know what's going on fast. And so this is where she can turn to copilot and just pull out again. Looking at all that feedback from your customers, can you summarize all the recent feedback on the landlord report export feature? And she can time box it. She can say, you know, in the last few days or the last 30 days or based on what priority customers have asked for, and you can even put out there saying, can you suggest some mitigation options? She's asked for it to come up with some mitigation options for you. So what you can do with AI is connect the dots between problems and solutions. You know, what she gets back is she can see a trend. There's 15 reports of failed CSV up downloads, there's mentions of blank files, there's complaints of timeouts. So it's flagging this all up. And it's not just that it's pulling out the problem, it can actually suggest mitigation paths too. So it's suggesting doing a hotfix, roll it back, or setting up manual CSV delivery for affected key customers, and can even write the comms for in app and email messaging to make sure that people know what's going on. So just from that instant of there's a big fire that the customers know about or that support team knows about, here's how you might go about handling it based on everything we know that your customers are saying to you right now. And the thing is, with AI, if you ask a vague question like what issues are there with landlord reports should probably be drowning in a whole bunch of irrelevant feedback. But you can narrow the scope. You can narrow the timeframe and the customer segment and jump straight to that particular set of signals. And to potential solutions as well. So really you can use AI to get past the noise so you can act fast and with confidence and just sort of quell these fires as they come up. So, all right, that escalation's under control. Ella is ready to shift gears, right? She's looking ahead. She's got a discovery interview with a major customer this afternoon and she's already drafted a rough outline of her discussion guide. But she's worried that her questions are leading. It's difficult to come up with questions that are going to get you information about how the customers are reacting without accidentally leading them to say what you hope they're going to say. And of course, if you have leading questions, you'll end up building stuff that you wanted to build rather than what your customers needed you to build. She knows that if she frames them the wrong way, the customer will just parrot back what she wants to hear and that's useless. So she's wondering, like, am I biasing the answers or the questions without realizing it. And so she asks, copilot, copilot, connect as a coach here, could you create a discovery guide for this customer over here focused on, you know, in this case it's on tenant communication area of the product. You know, use our vision, our past feedback to tailor it so she's able to get copilot to help her think about how she might start asking interesting questions that will get this customer talking. And here's what copilot gives her back. Right? So it's suggesting a warmup sort of question. How do you currently communicate with tenants? It's suggesting some deep dive questions. You know, what's most frustrating about managing tenant updates? It's suggesting follow up questions. You know, if you could wave a magic wand, what would this process look like? Right? So it's, it's giving a guide through how you might deal with this particular upcoming discovery question. Discovery interview. And of course, if you just pulled a generic guide off Google or even from an AI tool, you'd often end up with a bunch of irrelevant questions. Whereas this is very much tied into what do, what do we know about this product and this customer and this particular area that they're focus on focusing on. So it's a tailored contextual guide. Again, AI, the value of it is about how much context you're able to give it up front. And oftentimes that means writing a big long prompt, a lot of information that you give to it up front. Whereas with the copilot and prodpad, all that context is already fed in there. So you can just give it quick questions going, hey, I'm talking to this customer next. What should I say? So AI is great for prepping the guide, but it's really key that you bring the empathy you sense. Check it. And that you bring the humanized part of it and that you yourself run the discovery meeting yourself. You can't replace you in that meeting and having that conversation and listening into your customers. So interview prep goes smoothly. Ella's feeling more in control. It's going well after a tough morning. But then just after lunch, her CEO pops by with the. The dreaded question. You know, our competition's just done this. Can you tell me what's going on? Right? So our competitor, Buildium, what have they been up to recently? And of course she's sitting here going, oh God, I don't have time to pull together all of this competitor intel. I don't know what I'm not up to date on. You know what, where can I get this information quickly? So she asks Copilot, which has the ability to go off and search things for you, but also it knows what you are building as well. So it can flag potential opportunities and threats for your particular product. It can do comparisons between your two products. And so she's simply just throwing it in there saying, hey, Copilot, can you just give me some updates on this so I can feed back to my CEO and it delivers, right? So it's able to come back with a neat summary of Buildium updates. They've, they've redesigned their ux, there's some consolidated settings, they've improved their workflows, some opportunities that it's flagged. You know, you could differentiate on tenant communication, you could integrate faster. It's identified some threats that it brings as well. You know, the all in one appeal and some rising UX expectations. Now see, bad AI use would be to just repeat this blindly, right? Copy, paste, send to your CEO. That's not what Ella does, right? And basically that's how hallucinated product information and launches get get end up in exec meetings, right? What Ella does, because she's good at using AI, she fact checks her sources and she's able to see the sources that Copilot's come up with. She highlights what's confirmed and she flags anything that's uncertain and is able to put together something for her CEO rather quickly, given the, the scope of what they came by asking for earlier. So AI is fast, but remember that it's your judgment that that builds trust here. All Right. So Ella's day continues. She's seen enough fire drills for one day, but the chaos isn't over. Does this sound like anybody's day so far? Anybody having one of these days for this week? I'm seeing hands go up. Right. I appreciate everyone making time to join a session where we can actually learn about how we could speed up and get away from some of these fires. So yesterday, Ella ran a workshop. And of course, workshops are brilliant for pulling out information, but they're also great for creating a whole bunch of stuff that then needs to be organized. And she's like, I don't have time to make sense of all this mess. This is a lot of sticky notes. And so she asks Copilot, right? She's exported it into a PDF and she's just thrown it at copilot. And she said, can you just group these into initiatives, check for vision alignment, suggest quick wins and strategic bets. And she can ask it things like checking for alignment with the vision and with other parts of her backlog and her roadmap, because all of that is in one place, it's in prodpad. And so instantly Copilot Pilot is able to come back in and it's able to say, hey, yeah, people were talking about that simplify the. The onboarding problem. So suggesting to simplify that, that's very aligned with the vision. It says it's suggesting to expand landlord reporting. That's a medium term bet. It's suggesting a notifications overhaul as a quick win, right? So it's using this based on not just what it saw from those notes, but also what it can see in your own backlog, what's worked, what hasn't worked, what's your company aspiring to do. And again, you can't just accept everything at face value. You might miss nuance. So when using AI, use it as a first pass and then use your own instincts, your own intelligence to reorder based on the business context and the stuff that AI doesn't have its eyes on. Because remember, AI wasn't in that meeting yesterday. It wasn't in the leadership meeting that you joined the day before. So AI can really speed up this process and get you to the point where you can feel more confident making these decisions. So AI organizes, but it's up to you to prioritize. So it gets you from messy stickies to actionable initiatives in minutes. All right, so the chaos is under control again. She has space to reflect, and so she's going to pause and sense check one of her key Decisions, right. It's the turn of the quarter and she's got to think about the, the okrs that they're going to be tackling. And she's got a key OKR around, a key objective around onboarding. So she's wondering whether the key results that she's written for these are any good. Right. Are these even measuring the right outcomes? And so what she's doing is she's asking AI to sense check these key results, treating Copilot like a coach. And so Copilot delivers on that. Right. So Copilot responds here with informed insights around different metrics. You know, it's mentioning the, the heart metrics and the ARR funnel, the pirate metrics. It's outlining the, the strengths and the weaknesses. You know, the fact that she's got clear conversion targets but that it might not measure long term retention. So it's able to call you out on things as well. And a good AI should be primed to not just be a yes man, right? It should be primed to work as a coach, as a sparring partner. So AI is like a coach, it's not a decision maker. But what it's doing here is helping her to get clarity, letting her sense check things, having someone to, to spar with and to, to check whether her, her instincts are on the right path. And so it helps restore confidence. Right. There's nothing better than getting a sense that you're on the right path or tweaking something based on feedback and feeling like you are closer to where you needed to be. So with that, she's feeling like she's going to wrap up her day on a high note after all. But there's a new feature going live and it's happening tomorrow and she needs to close the loop with her beta testers. So she needs to make sure that everyone in her beta group knows that this functionalities coming out and that she wants to thank them for giving their feedback on it and asking them for additional feedback as it goes live. But she doesn't have the energy to write the type of email that's going to inspire confidence and inspire people to come and try this product out and all that sort of stuff. And so she can use AI again, right? She can say, hey, can you draft a friendly email in our brand voice and target it to our beta testers of the, in this case it's the Google Maps integration, pointing out how well it's been received and pointing out that it's going to general availability and thanking them for their feedback. And involvement. Again, this is something that it might just take 10, 15 minutes of your time to craft. But Copilot, your AI tool, can do this in a second, right? And you know, she might just send this as is, but Ella's smarter than that. She knows that she's going to add some, some additional love to it, right? So she wants to make sure that it feels human, make sure that it sounds just like her, that it represents the brand well. So it doesn't just ship something directly off to her customers. She takes it. She's the human in the loop here and she's able to improve on it. But it takes a job that would have been 15, 20 minutes and turns it into literally five minutes. And so with that, she's able to get out of there early, right? In the old days, she was drowning in admin and she was always scrambling for answers and she always had customers and stakeholders tapping her on the shoulder waiting for stuff. But in the new day, she's offloaded a lot of that bottom of the pyramid to AI. So she's confident, she's proactive, she's able to focus her time on the strategic work. She's keeping her customers in the loop and her stakeholders in the loop. And she was able to shut her laptop at 4:30 and get out there and enjoy the last of the the sun. So, you know, AI isn't about working harder or longer. It's about shifting your time from the admin work to the impact work and giving that admin work to AI. So there's still more that I want to cover here. Dive in a little bit deeper. But a lot of that I was talking about was around what Ella was prompting. Now, if you're using Copilot, as I said, Copilot already has all that context, but we have written a guide here. This is a free ebook that covers how to use AI to work better and faster. And this is a guide that has prompt information, like guides on how to write good prompts and give all that context regardless of where you're doing that work. So definitely grab this book if you are looking to craft your own prompts or if you want to speed it up, then you can try the the Copilot. We'll send you out a link to this to check out afterwards. But here's what you can't offload, right? So this is the top of the pyramid stuff. And I hope we never hit a point where we're trying to replace this work as, as product people, right? This is the stuff where we're actually adding value. Right. AI is powerful, but there's parts of product management that should just stay your hands. Anything where judgment, empathy, leadership, they all matter too much to delegate. Right? So think back to Ella's day. She used AI to draft her stakeholder email, but she was the one who knew which customer win to highlight. You know, she was the one who aligned things to the vision and used that vision to help guide a lot of the other information that she was using AI to pull through. So AI can remix, but it should never create your unique North Star, because AI is not particularly good at creating unique as we know. Right. It's good at creating the same as everything else. So when you're looking to create that unique, you've got to have that human leading the way. And that's where product managers are should be spending more and more of their time. Now that this busy work had been taken away, when she grouped those sticky notes from the workshop, AI gave her a head start, but it couldn't decide which initiative deserved priority. Right. She's got to make that final call. You can't offload key decisions to AI. I actually wrote about something like this similar the other day around the ethics of AI. You know, if you're offloading important key decisions to AI, who's responsible for it? Ultimately, you are. It just means that you've got less control over the output. So make sure you're still the one owning those decisions. You know, when Ella ran her discovery interview, she, she, she used AI to help her prep the questions. But while she was in the room, only Ella could follow those tangents and build trust with that customer. Right. Look that customer in the eye and, you know, show that she understood where they were coming from and dive into the areas that were really important. AI could help structure it and give her guides to go in there feeling prepared. But ultimately, you know, getting right in there in front of the customers is something that again, I hope we never lose touch with. You know, when her customer asked about competitors, she didn't just parrot what AI said, she fact checked, she filtered. She judged what mattered based on what she knows about those stakeholders. You know, so she's got to navigate that alignment and manage stakeholders herself. And when things broke, AI was able to surface the patterns. But Ella was the one who set the tone with her team, right? So she's got to lead the team and handle the crisis management herself. And you know, when she ended her day on getting that customer comms out, again, this is just something simple that AI can take, but it was Ella's human touch that made the email land, right? So she's still championing her customers, she's still being the voice inside the, inside the business. She's still guiding how customers hear from and those different touch points that the people interact with the company around. So she's offloading grunt work, but it is important to keep the human work, right? So vision, trade offs, relationships, leadership stuff that should all stay with you. But there are jobs that you can be offloading. So think about LSD again here. You know, in the strategy area, there's things that you can be offloading here. So you as a product manager, setting the strategy, but using AI to summarize it and communicate it in a way that makes it really easy for people to understand, right? You should be setting your own goals, but you can use AI to help sense, check, ask you questions about those goals, push back on them, craft, write them in more clear ways. You know, similarly with ideas, right? You can't use AI to just generate a whole bunch of ideas and just hit go, but you can use it to help brainstorm, right? Come up with a whole bunch of different areas that you might tackle so that you're not just thinking only about the one idea that you came up with, but you're seeing a wider picture when it comes to discovery, right? You can be using AI to do market and competitor research. Again, you got to sense check this stuff, user research again. It will help you dig up information and help you frame questions, but again, you can't replace the human part of that data analysis. Actually AI is brilliant for this, right? But do double check its work when you're having it do AI analysis and better to feed it smaller chunks of things rather than all of your data all at once and have it try to come up with something. So there's lots of ways that it can help you here, but lots of ways that you could end up over relying on on AI if you're not sense checking it. And AI prototyping is a really exciting area that we're seeing expand, right? I think we're going to hit a point soon where we'll quite literally be able to just say to an app on our phone saying, hey, can you make me an app that looks like this? Could you come up with a number of different ways that we might experiment with this? And as the product person, this isn't taking away from your job, this is enabling you to do more of your job, right? To dive deeper now instead of having questions or starting lines of communication around what's going to work for your business and testing out based on, you know, feedback forms, you can actually show people what it might look like and give them different variations and see how people interact with it. When it comes to feedback, AI can be really brilliant at helping to summarize and both individual pieces of feedback as well as whole bodies of feedback and analyzing what people are coming up with out of it and lots of other areas that can help with. Right, so stakeholder communication, right? Just taking your draft notes after a meeting and helping follow up with it with, you know, much clearer communications, coaching and best practice advice. Use this one with caution if you're just asking ChatGPT because it sort of reads everything from everywhere, right? It gets a lot of its information from Reddit, for example, which isn't known for its great product management copilot. On the other hand, we have primed it with all of the good practice stuff, right? So it's already coached with what good practice product management practice looks like. And so you can ask it questions and it will give you sourced answers on, you know, how the best companies do things. So there's lots of different areas that you can be offloading to AI to remove a lot of that grunt work. But I do want to just take a look at a few things that you should be considering, right? So you know, AI is really powerful, but only if you use it wisely. So LSD showed where some of these pitfalls could be right? So the yes man bias, for example, you know, if she just asked, are these okrs good? AI would have just nodded along. Instead she asked it to challenge her assumptions and that's when it became use, when it became more useful. And so, you know, some things that you can do here, you don't want AI to just be that yes man. You can tell it to disagree with you, right? Challenge my approach, find flaws in my reasoning. You can ask for multiple perspectives. You know, present this from a viewpoint of a customer, a competitor and an investor. Give me two arguments for this idea and two arguments against it, right? So we've actually primed our copilot to. To be a bit of a devil's advocate when it needs to be right. So it won't just follow along and be like, sure, that's the best idea you've ever come up with. It doesn't think that it will actually come back to you and say, oh, well, here's ways that you might improve it, here's some risks and challenges you might consider. And so it just gets it right there, right out there in front of you. But you can use AI to, you know, to become a bit more of this devil's advocate, right? So you can force it to rank and justify, you know, explain why one of these ideas is better than the others. You can do the reverse the burden of proof. So you can say, hey, this idea is terrible. What's wrong with it? Explain why this wouldn't work. Explain what could possibly go wrong with it. Make it poke holes in the evidence, right? Highlight anything missing or any weak assumptions. Again, we've already primed our copilot to already think along these lines because we can't assume that every product manager is coming up with all the brilliant ideas all the time, right? So we've primed copilot to be the, the devil's advocate. But here's some prompts you can use to really, really push in on this and remove that yes men bias. It's really important to give context, right? You only get the most relevant outputs when AI has as much context as possible about your product and your vision and your market and your customers more, right? So you've got to think about it as your AI tool is at best a junior member of your team who, if you give it information, it can scurry off and do some really interesting things for you. But the more they know, the more they contribute. And if that junior member is coming into certain meetings, if they were part of various conversations, then they can use that context. But remember that AI isn't in your meetings, right? It doesn't have all the context. And so there's going to be pieces of context that they miss. And so the more you can fill it in on what's in your head, the better your AI can serve you. And there's always a risk of hallucination, right? You know, if you're doing competitor research, you know, it could send you off in the wrong direction. So this is why you've got to sense check that stuff. In prodped's copilot, we've, we can, we've turned down the hallucination dial to basically zero, right? So it's, it's not a creative sparring partner in that sense. It's there to help connect the dots for you, point you in the right direction and ask some questions of you, right? But AI, different models work in different ways. Some are more creative than others. And so be aware of that and always sense check things, ask it for sources, ask it to, you know, point out where it got things and where things might, it might be getting Things wrong. And you can't just run, set up AI and just run, run with it, right? You're still the editor in chief, right? You can't just offload everything to AI and copy paste its answers and, you know, feed that into the next place and hope that it works, right? Ultimately, you're responsible for the quality of everything that you put your name on. AI is just a tool that you can use to get there faster. But you've got to be the one sense checking things. You've got to be the one who understands what it is that it has put forward to you. And if you're using it like a sparring partner, if you're using it like a coach, if you're using it like something that's going to help you speed up, but you're actually reading its stuff back, digesting it, then you're going to be able to deliver more and more valuable stuff as a product manager. But, you know, treat it like it is a junior PM on your team, right? It's useful. But don't publish without review because it's your name, your neck on the line at the end of the day. And remember that you've got to consider the tone of voice. And AI is getting good at learning your tone of voice and or adjusting tone of voice. But AI does tend to tend to lean towards having more of a overly formal, bland sort of language. So you need to actively shape its tone, right? How does your product personality need to show through? What's your own voice, especially when you think about anything customer friendly? So you can ask things like, you know, rewrite this in a friendly, confident tone, like a peer giving advice, or add a touch of humor and warmth, make it less cheesy, make it more cheesy, whatever you need to say there, right? And this, the beauty is that you can have it rewrite and rewrite and rewrite as much as you like. You know, ProgPad's copilot has no limits on it, so you can just keep playing with it and seeing how it works, seeing what works for you and then pulling out the best stuff. And that's the stuff that you can put your name on once you've read it through yourself. So I want you all to just take a moment and imagine your own day without the busy work. I mean, tell me in the comments, what would you guys be working on if you weren't working on that, all that busy work stuff, right? What would this feel like at the point that you've got everything, Admin, repetitive tasks, all that boring work offloaded. So you can actually focus your time on talking to stakeholders and learning from them and looking out at the market and putting it back into your own progression with the product. And I can see people putting in answers like, you know, thinking, just spending the time thinking and developing their, their thoughts, testing and validating ideas with customers. Right. We, we've always been told we know the stories to get out of the building, go talk to customers. And we've been so busy doing stuff like reshuffling stuff in our backlogs and having to write detailed specs that actually, honestly, we haven't had as much time to get out of the building. And so, yeah, loving the answers that we're seeing here. People wanting to spend more time being a better coach or mentor for other people. Product strategy work, spending more time one to one with stakeholders, you know, focusing on the strategy. There's so many interesting things we could really be doing when we get a chance to think and look beyond the horizon that's right in front of us, look beyond the mess that's on our desk. Jake here said testing, validating ideas for customers. And I think that's, that's like so, so crucial. One of the things like Janice said that AI, you know, shouldn't replace is that building relationship with customers and being at the more being able to spend more time actually with customers so that you know what you're building, you know, there is more and more evidence for what you're building and therefore you are building more and more of the, of the right stuff. Yeah, absolutely. I love that. So I encourage you all to jump into prodpad and try Copilot yourself. If you're not already doing that, you can access here, you can start a free trial or we also have a sandbox version prodpad that's preloaded with example roadmaps and ideas and experiments and feedback and all that sort of stuff. So you can actually play with Copilot within someone else's space. Right. The product there is actually apartment tracker, so you can go play like you were Ella for the day and see how it works for you. But also, as I said, there's a guidebook that we can send you that has tips on how to use AI, including doing all of those prompts. Right. Setting that context for AI too, so that it can be as effective as possible. So on that note, I hope we've all learned some things. Taking a look at how Ella has gone ahead to remove most of the grunt work from her day and is leaving early I'd love to hear your questions. Let's really dive into this. I'd love to hear your comments, see what sort of things resonated the most with you and any questions you've got about how this might apply to your own work. So we've so, so there's still time to add any questions to that Q and A box and we'll, we'll go through them now. We've got a bunch in there at the moment we can start to tackle. Some are some more theoretical, others are specifically focused on on copilot. So the first one Michelle has asked how does understand the tone of the user? So I think this is Jana, when you were talking about asking Copilot to draft a stakeholder email and how it understands the tone of the user. I will just say quickly before we answer that, that unlike chat, GPT or other general tones, you know, the, the team behind Co Pilot has spent a lot of time fine tuning and priming Copilot. So and Co Pilot tone of voice straight out the box is, is friendlier than your, your traditional and general AI tool. So even without the in my usual tone, you'll get something which, which is friendly yet professional, I would say. So it understands what you mean by customer comms. It understands, you know, where how you might want to talk to an exec, how you might want to talk to an internal team, how you might want to talk to customers. But also like other AI tools, it's able to learn about you. Right? So our Copilot, what it is, it's basically a, an AI that has access to your entire product backlog and your, your, your vision and your roadmaps and things like that in a tool prodpad that everyone in your company has access to. So everyone in your company has access to this, this copilot. But when they're talking to it, it will learn their behaviors versus your behaviors. So it will pick up on your tone. Now one of the things we talk about in that ebook is how you can set that context. So if there's a particular tone that your brand has or that you have, give it examples saying here's what we did for last month's release. Can you make one for this month's release? And the next time you go and work with it, you won't have to give it that context because it already knows what your past releases look like. But with AI, it's always about giving that context and giving it guidance on, you know, what it is that you're looking for. And if you Find that it hasn't hit the first. The tone the first time. Just give it that tweak and say, hey, so like, I've done that, but can you please do it more like this? Right. Can you be more formal or less. Yeah. Giddy or. More giddy or whatever. Or change it from, you know, I do this all the time. Change it from American English to British English to. I use Canadian naturally, because that's where I'm from. So. So yeah, it allows you to. To change it up. Great. The next one is really quite technical from Bradley, so more of a technical jumping around, more of a technical question as we're looking at building some functionality like this. So, like into there, into Bradley's CMS product digital, you make use of MCP to pass context to Copilot in some. Places, but not as much as it will do in the future. So Copilot is a new product in beta and so we're making use of some of that, but there's more that we're planning on doing in that. So definitely, definitely keep a watch on our new releases here at Prodpad. We do releases every week we have done for the past decade, so there's always new stuff in there. And what we're really doing, building for AI is actually really interesting right now because it's a little bit like building for mobile back in the late 2000s. Right. Like, people are still figuring out what, how it works and how people interact with it. That's where we are right now. Right. We're right at the forefront and figuring out, you know, what do people expect from a chatbot qa, Sorry, a chatbot sort of system? What do people expect when they press the purple sparkly button? And everyone knows the purpley sparkly button means AI now, but what are the other patterns that are going to develop over the next few years that are going to set how we interact with. With AI? So, you know, watch a space as this evolves. We're product people building tools for product people. So you can imagine that this stuff goes through a lot of iterations as we learn. But importantly, Copilot in terms of, in. In terms of getting context. Copilot is a feature within prodpad and so it is built within the context. So, so Copilot has access because, you know, it has access externally through APIs, but it has the, the internal context of your product vision, your roadmap, everything in your POD account, because it is a. It is a feature within, within popad. Yeah, excellent, Excellent. Shall we? So there's Another question from Kim. Yeah. Kim asks what AI model is Copilot using? So again, this one's getting technical into how we're doing things. Right now we're using GPT 4.4.1. 4.1, yeah. To be exact way we've built it is so that we can switch it out for other models as and when they become. So right now we're using OpenAI, we might switch to other models. These models are all coming through fast and hot right now. We're currently testing out five and making the decision as to whether we switch to that one. We're looking at other places. So basically we're, we're building it so that it can be model agnostic and we're going to choose the one that is most powerful and most useful for the use case that we have, which is for product people and their teams trying to build the best products. And then Kim, the other half of your question there was with will the information that you give Copilot feed to train the model or would it be contained? Yeah. So luckily it is a. Sorry. Very importantly, very importantly, we do not send this training data or any of the stuff that you use within COPILOT into anyone's training data. So no, it does not go back to OpenAI. No, it will not go back to any other service provider. And your copilot within your company is self contained. So none of this stuff is shared between different companies using different instances of copilot. So I think that's really important. We're very, very conscious about security and the privacy of your data and it's. A closed environment, which actually relates to John's question. So John has asked, do you have any advice for persuading employers who are wary about AI and haven't integrated it into our toolbox? So some of that will go, you know, to that. If you use the right AI tool, it is secure, it is, it is closed. You don't have that risk. If one of the, one of the concerns there is about the risk of sor, of data being used to feed broader models. Also when it comes down to persuading your employers, you've got to think about what the different objections might be. You know, is it security? Is it that they don't believe that there's benefits? So each one of these different things that you might tackle might have a different response to it. You know, sometimes it's one of these things where frankly it's easier to ask for forgiveness than permission. I know a lot of product people who are using like using AI tools in general and their bosses don't know, their teams don't know. But I think what's going to happen is because it's going to become such a common use case and people are going to become much more effective using it, that'll become the norm. Right? It'll become accepted. But there's always going to be companies who are early adopters and always going to be companies who are late adopters on this stuff. And, you know, different teams might need different levels of confidence in it before they get there. So if you're having your company push back on it, ask why, figure out what's really stopping them from doing so. Is it inertia? Is it fear, is it anxiety? What is it? And then, you know, you can start figuring out how you might gently pull them into the future. I also think that sort of value hierarchy, the pyramid that Jana showed right at the start of the webinar would, is really compelling and could help with the, with the argument when you show. And maybe, you know, maybe you go through the exercise of, of illustrating or quantifying where your team and yourself are spending the majority of their time and suggesting that the introduction of AI tool will, in that enable you to spend more time at the top of the pyramid and therefore deliver more value. Yeah. All right, great. Okay, we've got time for a couple more questions, so let's keep jumping into them. Jay asked, do you have any best practice advice around data quality and Data needs within Prodpad? Specifically around what CoPilot uses within Prodpad to ensure the prompts you've gone ahead are you've gone through are realized. So prodpad, what it does, it structures this set of information around your product processes and your product information, right? So it has specific places to capture things like your vision and your okrs. You know, which of those OKRs have worked, which ones haven't. You know, you can see feedback from your customers. That's all structured. It can see ideas, it can see how ideas are connected to the roadmap, your big strategic steps. So all of this stuff is quite structured. You know, at a bare minimum, we generally recommend that you have the vision area, the, the, you know, your product vision and high level information about your products filled in just to get started. Because that gives it context as to what it is you're actually trying to build towards. Right. We even have a tool that'll actually help you write your vision. Right? You can give it a basic version and then you can say, hey, can you judge this? And It'll give you some, some feedback to improve your vision until you get something that's going to be useful for both you as well as your team and the AI. So basically, the more like with any AI, the more context you give it, the better, right? And this is the type of tool that's going to get better and better and better over years, right? We're working with people who have been using Prodpad for like a decade, right? They've got years and years worth of decisions logged in Prodpad. And that's really interesting context to magically switch AI AI on for, because it can look back and say, hey, here's stuff that you did in the past. Here's. Here's what worked with it, here's what didn't work, here's what the team discussed around that point in time. So some really interesting things that you can do as you capture more and more information in there. And I think going forward, as AI gets more and more powerful, powerful will be, you know, game changing for, for companies. And I think, yeah, I think certainly, like Janice says, the sort of baseline is to make sure there's. There's an area in proper called the strategy Canvas where you put in, as Jenna says, your product vision and that is fully customizable, so you can put whatever sections you want. So you might, you could, you could add some. A deeper description of your product etc. But one of the, one of the cool things that that copilot has done from day one is actually a bit of promptless functionality where every time anyone submits an idea into projpad, there is a button to hit that. That for alignment assessment against your product vision. So if you've got a product vision in there, this is great with stakeholders, particularly those stakeholders. This is going to sound awful that, that have that submit sort of ideas that are like, that's. We're not going to do that. That's rubbish. So you can be a winner. No, but this button will give them that direct feedback so you don't have to. And it'll say, this is not. This is. This is not a good fit to the vision. Here is why. And it'll give suggestions for improvements. But anyway, so I just wanted to. That that's a particularly cool thing. It's been very popular with our customers. This. Excellent, Excellent. All right, so we're running low on time here. We've got a few more questions in here, but I'm going to pick one out because we've got somebody asking about how does ProdPad compare to its competitors. From an integrated AI perspective, what are some of the conversations or arguments that would help me steer the conversation towards Prodpad versus other players? I mean, more than anything, jump in there and try it for yourself, right? What we're hearing from people is that our AI is miles ahead of the competition. You know, we've been doing, we've been playing with AI, so since before it was cool, we had a version of a chatbot that we called dotbot that was based on gpt2, so predated chatgpt, and it allowed you to ask questions and talk to it about pieces of feedback in your backlog. We've upgraded that and upgraded that and upgraded that and now given it access not just to your feedback, but to your entire history of all those decisions, your entire vision, your goals, your, your customer comms, your everything. Right? So this thing is now just really, really powered up. We've got a head start on this. We're miles ahead and we're also, you know, as a bootstrapped, private, profitable company here to take care of you as a customer. Your data, we're not going to be selling it off to, you know, sending it back for training data or doing untoward things with it. We're here because we're the company that actually helps you build better product management with better tools and helps you do better product management. So we're really proud of what we've built. There's so much that you can do with AI in general. There's a lot so much that you can do with AI built into Prodpad. We've got guides on both of those, so we'll follow up with those afterwards. In the meantime, we're at time, so I just want to say a huge thank you for everybody for joining all the questions. I'm really sorry we didn't get a chance to go to all of them, but if anybody has any follow up questions, shoot us an email, reach out to us or jump into the product itself, start a free prog pad trial or jump into the sandbox that Prodpad has and just start playing around, letting us know, let you know, reach out and let us know what you discover with it. So once again, thanks so much. It's been a great session. We'll chat to you again next time. Thanks everyone. Bye. All right, bye for now.

Head Shot

Ben Maynard

Host