# AI Emptied Our Backlog. Now What?

Verena Graber, Just Product 2026, 10:30. Automatic transcript of the livestream (Parakeet ASR, unedited, expect misheard names). Timestamps are positions in the talk video.

[0:01] Alright.
[0:02] So next up, we are entering the AI mindset theme.
[0:07] So I have to also tell you a little personal story about the next speaker.
[0:14] And maybe we can just show her on the screen here.
[0:17] So that is Virina.
[0:19] I met Verena like I think beginning of the year, and she was uh participating in one of our cohorts, and she said she's feeling like an octopus with 800 arms.
[0:29] And that stuck with me because I think like every PM can relate, it's sometimes just a lot.
[0:36] And then through some iterations, she will talk about in a minute, she somehow miraculously managed, he she and her team to completely clean out their backlog.
[0:49] And I think that is something that is basically unheard of.
[0:57] So if AI is emptying our backlogs, now what?
[1:01] What are we doing now?
[1:03] So please welcome on stage Verena Gaba.
[1:18] So I reduced 800 to 100, but this story begins with me feeling like an octopus with 100 or 800 arms.
[1:29] So many I couldn't even count anymore.
[1:32] You all know product is an incredibly diverse role, especially in a market that's grown so fast, products that are evolving at lightning speed and always changing technology technologies, features, and tools.
[1:50] So while running our product department on my own for a while, my work spanned from research, ideation, design, roadmaps, collaborating with developers, testing, documentation, stakeholder management, etc.
[2:12] etc.
[2:13] And because the product evolved so quickly, I also helped onboard individual clients until the day they go live and joined sales calls to present our newest additions.
[2:46] My name is Verina, I'm the product manager at MadFlex.
[2:49] MatFlex is a tool that supports.
[2:52] AI supports communication and structures and automate processes for healthcare providers.
[3:00] In this talk, I share stories about how I used AI to speed up tasks that kept falling through the cracks, and how this led to an almost empty backlog.
[3:17] And how product was the bottleneck again and again.
[3:23] To be honest, today I still feel like an octopus sometimes, and maybe that's needed for our role.
[3:32] But my arms are much stronger now and better controlled.
[3:40] As you heard, I've been to the product management master class, the product management in the AI era class in spring this year.
[3:49] And this class gave me the chance to step back from my busy day-to-day work.
[3:55] I was talking to like-minded people, could experience new techniques and tools, and was had these three hours a week blocked just to experiment.
[4:43] Nothing will ever get built.
[4:46] I had to step back from what was missing and focused on what was already there.
[4:53] I mean, I had the experience from these eight weeks.
[4:55] What could I was able to do with very rare footage and very limited time?
[5:02] So that's what I did.
[5:05] Sales and customer support already transcribe most of their calls at our team.
[5:11] So I used these transcripts and put them into two folders.
[5:14] One was for new clients and one was for already existing ones.
[5:21] I did that because AI keep mixing up what already exists and what was new.
[5:27] So we had weird ideas.
[5:32] Next, I took AI to go through these transcripts and find out what was it that customers complained about all the time, what new customers needed from our product, and what questions were asked the most.
[5:47] Later, we also added the support ticket pipeline to find out which bugs kept occurring and what we needed to fix right now.
[5:59] This is still happening, so every new transcript that is existing is put in one of these folders.
[6:06] They kept keep being scanned by AI.
[6:10] I get the list of ideas and inputs, and I check this against our backlog.
[6:17] So it's an it turned to a living process with almost no human interaction.
[6:24] So next, I took all the insights I got from this process and challenged it against my backlog.
[6:31] And I don't know about your backlog, but ours was about five kilometers long with tickets from years ago nobody remembered about.
[6:40] I found a ticket from two years ago that was got somehow got lost in the list, and it was still relevant today, and it was the first thing I put up for ready to pick.
[6:50] And another thorn that I just created last week got deleted immediately.
[6:59] So this easy process, um, scanning our backlog and using the input that I already had.
[7:07] I didn't need to spend hours and hours on talks or hours recruiting the best clients to interview.
[7:17] Um helped me to almost reduce our backlog down by 50%.
[7:22] And it assured me that what was ready for development was needed and recently reviewed.
[7:30] So nothing was forgotten anymore, and everyone knew that they worked on relevant stuff.
[7:40] It was fun, but the tech team turned in, and they introduced agentic engineering.
[7:52] Um got assisted by AI agents that speeded up their development cycles by four times almost.
[8:10] It started on Monday.
[8:11] I've had a great list of to-dos for our developers, and by Friday nothing was left.
[8:18] And my Teams chat looked like you can see it here.
[8:21] I constantly was asked, what can I do next?
[8:24] I need a new ticket.
[8:26] Is there anything I can do now?
[8:28] Are you done with testing now?
[8:31] So I could see my beautiful new backlog shrink right before my eyes.
[8:36] And I panicked because there was just nothing left that I could feed development with.
[8:44] So it turned out product kind of is the bottleneck again.
[8:52] And AI development outrun product.
[8:59] So we were so busy testing and finalizing releases and keeping the team up to date that we had no chance to build new stuff or create new ideas.
[9:13] So we had to change something.
[9:15] And we looked at our methods.
[9:20] Like many of you, I think, we worked in sprints, and we killed that.
[9:27] Because what's the point of planning a two-week block for development when I'm mostly busy putting in tasks that are getting done within minutes?
[9:40] And thank you.
[9:44] Next thing we killed was story points.
[9:48] That was a little bit harder for me, I have to agree.
[9:51] I have to say.
[9:55] Our developers said story points make no sense anymore, and I knew they were right.
[10:01] But story points were just little numbers that helped me shape my arguments for my stakeholders.
[10:07] I knew how long something would take.
[10:09] I knew how many tasks I can plan to be done parallel.
[10:14] I had a chance to say no to things that didn't really match our path, and could explain this by this is just too big or too expensive to do.
[10:26] The outcome is not big enough.
[10:32] It took us no development, no.
[10:35] Development kept showing me that story points make no sense, and that I have to get rid of that now.
[10:44] I will use Scrum Poker, and I don't know if you know the interface, but you can estimate numbers or symbols, except for example, the coffee pot.
[10:55] And our estimations that before looked like pretty little numbers for every task now looked like a cafe in the middle of a city because it coffee pots kept popping up all the time.
[11:11] Developers were just not able to estimate how long something would take or how much effort it would be because the agent is doing it anyway.
[11:22] So one thing we kept was groomings.
[11:38] And we stopped looking at tickets and comparing them.
[11:41] If this ticket is a two, then this one has to be a five.
[11:45] Is this sentence in this ticket really correct?
[11:48] The design looks a little off.
[11:50] Can we change two pixels here and there?
[11:52] This took a lot of time, and for Stefan, my product designer who's also here today, and me, it wasn't always fun.
[12:01] So now we change what we look at when we go into groomings.
[12:05] We try to focus more on what is the goal of this idea.
[12:10] What can we do as a team?
[12:12] And the team means design plus development and everyone involved.
[12:18] How can we reach the goal faster, more easily, or how can we even improve what we think of here?
[12:49] And again, we felt that there
[12:52] The bottle the bottleneck is not solved.
[12:54] It's in our own methods.
[13:00] So we organized a meeting.
[13:02] Our CEO, me, and our two lead developers.
[13:07] And we looked at this prototype here.
[13:09] It's a rush sketch.
[13:10] Our CEO did, we are all visual syncers, in I think two hours while he talks to our most important clients.
[13:21] It held our vision and big ideas.
[13:25] Ideas so big we weren't even able to touch until now because we didn't know where to start.
[13:59] And then we gave it to the top three ideas to three developers.
[14:05] So one idea to one developer.
[14:08] They only had the rough sketch, the prototype, and the task, what can you do with this using your agents within two weeks.
[14:28] And they didn't impact the core of our product too much.
[14:32] So they could live on their own for quite a while.
[14:38] And developers starting doing that.
[14:40] And it was kind of a blind spot for us.
[14:43] And to be honest, it made me a little nervous because I know our developers are great, but they have their own personalities, and you can't expect what will come out from one or the other person.
[14:58] And what we planned on will take us about two years at the beginning of this year.
[15:06] Took two weeks.
[15:07] And all these elements are now live on our testing stages and being reviewed by our designers.
[15:16] So we took a process that would have taken us with our usual methods, and we just skipped most parts of it and worked together as a product and development team, really hand in hand, using all the agentic help we could have to get things done that we wanted to start for months, but really didn't know how.
[16:22] It has to survive this check.
[16:25] What we humans do is still decide what gets to be done and when.
[16:32] The second arm is for exploration and design.
[16:37] This is a part where humans are involved most right now.
[16:42] We still do designs using Cloud Design and Figma.
[16:48] We still discuss pixel detailed stuff with our developers.
[16:54] But we speed it up by getting faster prototypes and more variants to decide on.
[17:06] The third arm is for grooming.
[17:08] We don't estimate tickets anymore.
[17:10] We have killed sprints and story points.
[17:15] But we ask what is the actual goal?
[17:17] What would it take to make us faster or make the product more lovable?
[17:24] And how could it be done even better with us working together without having to write detailed tickets?
[17:32] That's also not a human job anymore.
[17:35] And I agent does that.
[17:40] And documentation was the part I thought was the most boring thing.
[17:44] I hated doing that.
[17:47] But what's cool now is that I'm able to talk about things we already can do now that is life today, and not about what I wish we could do sometime in 20 years.
[18:06] More important than it has been ever, because we're growing even faster and changing our product and the way we work so fast that we have to take the rest of the team with us, especially customer support.
[18:23] So I think there's a point that we can improve, but we're getting there.
[18:28] Very good.
[18:29] So AI agent write documentations and instructions that I still have to go over.
[18:36] That's the not so fun part.
[18:39] Um we're more training our team how to use these new resources instead of answering the same question again and again and again.
[18:53] What I want to share with you from this story is don't be afraid to be the bottleneck.
[18:59] Really challenge it to go to this point and then use it as a chance.
[19:06] Because what it showed us is what do we still need?
[19:10] What can we let go of, and where can we go work closer together to really reach our goals?
[19:20] So use it to define what you really need and what could be solved smarter.
[19:26] You don't have to find the perfect new method right away, we don't have that in place.
[19:34] Um making your role a little more fluid and be open to give things that you thought you were responsible of away to others.
[19:49] Thank you so much.
[19:58] Thank you.
[19:59] Wow.
[20:00] I really love that story.
[20:02] Um, I don't know where the box is.
[20:06] Is there do we still have that up there?
[20:10] Do you have any questions for Verena?
[20:13] Okay.
[20:14] Just need to find the box.
[20:16] Who has the box?
[20:20] Ah, over there, over there.
[20:22] All right.
[20:22] So we had a question in the front there.
[20:25] So all right, so the box is in the back.
[20:27] Turn around again.
[20:28] Safety first.
[20:29] Oh, sorry.
[20:30] That then this is the next question.
[20:33] Safety first.
[20:35] Hand it over.
[20:37] Thank you.
[20:37] Uh thanks Verena.
[20:39] Um thanks for your keynote.
[20:40] Uh uh, especially like the metaphor of the octopus with the eight arms for product management, because it's exactly that.
[20:46] Um, my question is if you eliminated sprints, what did you replace them with?
[20:51] Because a sprint is kind of uh like the the opportunity also to have a checkpoint at the end of a define time frame where you say, Okay, what did we do?
[21:00] How did we do well?
[21:01] Did we meet our goals?
[21:02] How do we improve for the next time to retrospect and improve things?
[21:06] If do you still have those checkpoints in a way?
[21:11] Yeah, thank you for that question.
[21:12] So the octopus is internally called the productipus.
[21:18] And yes, that was something I wasn't so sure about how to solve it as well when we uh decided to eliminate sprints.
[21:26] Um we kind of went back to Kanban.
[21:30] Um, but what we do differently now is that in the groomings we also discuss um when we need something or how much time it will give us to release that thing.
[21:42] And we found that we're um too fast with releases anyway, because our clients couldn't keep up.
[21:48] Um so we're in a lucky phase right now where we build stuff and wait for the release.
[21:54] So we do not really have that problem of checking in if something gets done.
[22:00] And we still do dailies.
[22:02] So if a ticket pops up in a daily for more than three weeks, we have a marker that something is off here, and try to discuss how to continue with that.
[22:13] Um yeah, but we really haven't figured out the perfect new plan to um make sure timing is correct.
[22:23] It wasn't it wasn't much about timing, but rather also about the collaboration part with yeah, mostly dailies.
[22:31] Yeah, mostly dailies, and we're a small team.
[22:34] Um most of us sit in the same office, so we talk a lot.
[22:39] Okay, thank you.
[22:41] All right, and there was I think another question in the front over there, yeah.
[22:47] Right.
[22:47] You can also throw it, don't be afraid, just everybody is.
[22:52] Keeping the eyes on the on the catchbox.
[22:56] Good.
[22:57] Is it working?
[22:58] Perfect.
[23:00] Based on your last lesson, you mentioned something about delegation to your team.
[23:05] Yeah.
[23:05] How has your team managed it?
[23:07] Was it eager to get those new responsibilities or a little bit defensive?
[23:13] No, they were really eager to start doing that.
[23:16] I mean, we all talk about involving development early on.
[23:21] And I always said that too, but I have to be honest, I didn't do that that well.
[23:28] So now we're forced to do that.
[23:30] And we have developers in the very early stages of ideation and design with us.
[23:37] And I think we're lucky that we have a shared vision of how our product should work and what's important to us as a team.
[23:48] So yeah, they were very happy to be designers as well, or to get to experiment a little more.
[23:56] I'm curious how it will be when we go through the um through the testing stages now because there will be a lot that needs to be reworked, and also design choices they made will be deleted.
[24:11] Um but yeah, I think with clear communication and always focus on the goal, that will work out.
[24:20] Is it working?
[24:22] Um just a follow-up and opass.
[24:24] Um just to understand.
[24:26] So, in terms of vision, was it the challenge?
[24:30] Do you feel that, for example, as your product manager you still have some control of okay, we have an objective and we're not spinning off into something completely different?
[24:40] Yeah, that's that's not so easy to do because um, yeah, when you can build everything, you have to um focus on what you really want to build, and that's uh that's a task I I try to work on every day.
[24:54] Um but right now, yes, we're very clear about that, and still we I have enough contact to sales to support to our customers to really find out if that what we planned is what we need, um and we're fluid enough to change that real quick.
[25:15] All right.
[25:16] I think we have time for one more in the front.
[25:20] Yeah, we definitely have two.
[25:24] In the front.
[25:28] Hi, thanks.
[25:29] Um in arm two, you mentioned the design and exploration.
[25:34] Uh how did you speed up the process of the validation with the customer?
[25:38] Because I have a challenge in that requires a lot of time going back and forth, back and forth, back and forth.
[25:45] Yeah.
[25:46] Um we do less of it, to be honest.
[25:51] And we test against what we already have and use our power users as a feedback giver.
[25:58] And we're very involved in the implementation stage.
[26:01] So we are able to talk to clients while they're implementing our existing product and um show them what will be done next and how this could even improve their workflows.
[26:16] So we're kind of um collecting feedback while the customer is in onboarding stage, and which helps us to find out if the stuff that will be live in the next couple of weeks really supports what they need and if they understand how it should work.
[26:33] So we have we use the the learning phase of a customer where they're still curious and ask a lot of questions for our advantage to get feedback very quick and we skip the um the uh scheduling appointments and finding someone to talk to through this as well.
[26:54] Okay, thanks.
[26:56] Perfect.
[26:58] Thank you so much, Verina, for sharing your story.
[27:00] Please give a big more.
[27:05] Yeah, that's it.
[27:06] Thank you so much.
[27:08] Oh, I'm gonna still.
[27:12] All right.