Sign in to view Christines full profile
Christine can introduce you to 10+ people at Fin
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIns User Agreement, Privacy Policy, and Cookie Policy.
Christine Zdelar
Sign in to view Christines full profile
Christine can introduce you to 10+ people at Fin
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIns User Agreement, Privacy Policy, and Cookie Policy.
Dublin, County Dublin, Ireland
Contact Info
Sign in to view Christines full profile
Christine can introduce you to 10+ people at Fin
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIns User Agreement, Privacy Policy, and Cookie Policy.
3K followers
500+ connections
Sign in to view Christines full profile
Christine can introduce you to 10+ people at Fin
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIns User Agreement, Privacy Policy, and Cookie Policy.
3K followers
500+ connections
See your mutual connections
View mutual connections with Christine
Christine can introduce you to 10+ people at Fin
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIns User Agreement, Privacy Policy, and Cookie Policy.
View mutual connections with Christine
Christine can introduce you to 10+ people at Fin
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIns User Agreement, Privacy Policy, and Cookie Policy.
Join to view profile
Message
Sign in to view Christines full profile
Christine can introduce you to 10+ people at Fin
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIns User Agreement, Privacy Policy, and Cookie Policy.
Articles by Christine
-
3 lessons from my first 3 months as GPM
3 lessons from my first 3 months as GPM
This week marks 3 months in the role of Group Product Manager. I spent the last 3 day weekend reflecting on the
705 Comments -
How to win the RESTART Marathon by putting PEOPLE FIRST (Part 4: Tortoise, Hare, & HR)Jun 4, 2020
How to win the RESTART Marathon by putting PEOPLE FIRST (Part 4: Tortoise, Hare, & HR)
The Tortoise, the Hare, and HR: Who will win the RESTART RACE? Peter R. Classen, Christine Zdelar - June 3rd, 2020 Part
14 -
Why your team needs to start every morning with a standupMay 26, 2020
Why your team needs to start every morning with a standup
Many teams are facing remote work for the first time due to COVID-19. For some, the last few weeks have been riddled
232 Comments
Activity
3K followers
- Posts
- Comments
- Reactions
No more previous content
-
Christine Zdelar posted thisI worry about a lot of product leaders right now. To lead well in AI (think: strategy, hiring, knowing where to give feedback, knowing what good looks like and why) you need to have been in it yourself. Not one or two layers of abstraction away from it. And many aren't. Here's what I've observed over the last eighteen months, having oscillated between two very different modes of working. The first half of last year I was deep in hands-on building with Fins Insights product. Embedded with the team, in the weeds, learning what it actually takes to build AI products from the inside. The second half I spent almost entirely market-facing in the recording studio, appearing on podcasts, sharpening the narrative around Fins new set of AI-powered capabilities, trying to get customers to see the value that this technology had unlocked for them. Now Ive swung back into deep hands-on building again with the onboarding agent experiment. What I've come to believe is that right now, as a product leader (GPM, Director, VP), you need to be able to do at least one of two things. Ideally both. Build. Not oversee. Not review. Not be in the room while others build. Actually get your hands dirty building production-ready AI products. You have to understand what it takes, what breaks, what surprises you, what your gut gets right and wrong. Sell. And I don't mean we all need a quota attached to our names. I mean get really, really good at communicating the value of what this technology has unlocked, directly to the market, from firsthand experience. Not from layers of abstraction away. The market is saturated with hype and vaporware. Your customers are being spoken to by a hundred companies making AI claims right now. Cutting through that takes a kind of credibility and finesse that you can only develop by being close to it yourself. The other PM leadership stuff like group level strategy, org design, people management has become less sufficient on its own. At least at Fin, the thing I keep coming back to is this: how do you know what good looks like if you haven't experienced it firsthand? You can read about it. You can be briefed on it. You can tinker on the side and vibe code a few things. But that is so far from what it actually takes to build something production ready. You can't borrow someone else's conviction or intuition. And so if you can't speak credibly to what it actually takes to build these products, or speak credibly to the market about what they can do... it isn't clear what your value add is. I say this having been that person at points in the last two years. It's not a comfortable thing to sit with. Reminder: none of us have done this before. We're all figuring it out. Might as well do it out loud.
-
Christine Zdelar posted this"Why haven't we shipped anything yet?" That was the message I heard from leadership about eight weeks into our onboarding agent experiment. *jump scare* My negative self-talk became loud and my ego was... not ok. What I *heard* was: "Why is Christine allowing this to move so slowly? Does she not know what she's doing?" So I did what many a PM would do when they need to show progress... I committed to a ship date. In the traditional world of software building, that would have been the right move. I've even heard a leader say something to the affect of "as a PM, nothing really matters if you don't ultimately ship something." Once you've figured out what you want to build, it's just a function of time, resources, and execution. It's always been *possible* to build the thing. The question is just how much it's going to take. With AI products, I knew it was different. I knew that *intellectually*. I had people telling me, "Right now, the thing you need to do is learn." But...I was in full lizard brain mode. So I still made the mistake. The real question we needed to answer wasn't "when can we ship this?" It was "can we even build a quality agent at all?" That question of feasibility doesn't really exist in the same way with deterministic software. With AI... especially zero to one, greenfield, generative AI... "can we?" is one of the fundamental first question syou have to answer. What I should have committed to was a learning milestone, not a ship-to-production milestone. Something like: "By this date, we want to know if it's possible to build a quality agent, and what it would take to get there." That framing would have shown more incremental progress. It would have given us flexibility. It would have made all the learning we were doing visible and made it easier to communicate the value of that learning. Instead, we had a ship-to-production date. And we didn't hit it. Then we didn't hit it again... It looked bad, and it certainly didn't do anything to help my confidence. Even though throughout all of that we were learning a ton. Which really was the whole point. I think this is one of the hardest reframes for product leaders right now, myself absolutely included. If you're a Director or VP, you've probably been less hands-on with building lately. Your own leadership might not be close to this technology either. So everyone defaults to the language and metrics they already know: ship dates, launch milestones, production. It looks like progress. But early on in AI product building, especially agent building, your job isn't to ship. It's to prove what's possible and learn what it takes. Those are completely different milestones that look and feel different. The way you measure and communicate them has to be different too. I had to learn that the hard way. As a reminder: none of us have done this before. We're all figuring it out. Might as well do it out loud.
-
Christine Zdelar posted this"Almost there" is a trap. Especially in AI product building. We had a small, senior team working on our AI onboarding agent experiment. The agent was working, powered by a playbook and a new front end Messenger experience. And every single day for almost three weeks, we'd hammer away at it with test conversations and think: Okay. Today's the day. This feels good. Let's ship tomorrow. And then we'd find something. A response that didn't land right. A plan the agent generated that was... off. Nothing catastrophic, just not meeting our bar. So we'd tweak, retest, and convince ourselves we were almost there. "Almost there" kept us doing the same thing every day expecting different results. It was my gut, eventually, that said: okay. This isn't working. We need to stop and reset. So we did something that felt almost *too* simple. We got a tiny group together comprised of me, our lead AI engineer, and our principal designer. I said: before we test anything else, let's just write down what good actually looks like. What are the qualities of a genuinely great onboarding conversation with this agent? What does a strong deployment plan look like? What's non-negotiable vs. nice to have? We'd never written any of that down. And here's the thing... I don't think we could have on day one. We needed those weeks of messy, scattershot testing to develop enough intuition about how the agent actually behaved before we could articulate what we wanted from it. The testing wasn't wasted. It was the research. Once we had that written down, we structured our evals around it. Nothing overly sophisticated, not some complex framework. We just needed a clear set of criteria, tested methodically, assessed individually. Within a few days we'd made the adjustments we needed. Within a week we felt genuinely ready to ship. In hindsight, could we have gotten there a week and a half earlier? Probably. But would I spend those weeks to learn this lesson properly? Yes..because I won't make this mistake again (I'll make others instead). The tricky thing is that this was a senior team. People I'd put on the "strongest AI thinkers I know" list. And we still didn't clock it sooner. That's not a failure. That's just what building new things actually feels like. Reminder: None of us have done this before. We're all figuring it out. Might as well do it out loud.
-
Christine Zdelar posted thisI've been a PM for almost 15 years, and lately, there are many days where it feels like.. I have no idea what I'm doing right now! And I think that's fine. Actually, I think it's the point. Yesterday, something happened that made me weirdly... emotional? We've been building an AI agent one that helps customers get set up and live with Fin for Service. Think of it as an agent helping people deploy an agent for the first time. It's an experiment that we shipped two weeks ago and held our breath. Yesterday, our very first customer who'd been working through their whole setup journey with this agent went live. It worked! Not perfectly. Not automatically. There was actually a moment where the customer just... went quiet. Got busy, got distracted, the thread went cold for a few days. We knew we'd need to build something to re-engage people at that point. We just hadn't built it yet. But I knew this customer was very close to going live. They had so much meanigful back and forth for over a week. My gut told me to get them back into the conversation and give it one more shot. So I faked it. I manually read back through the whole conversation, saw they were almost there, and sent them a personal nudge in the Messenger. Essentially said something like "Hey looks like you're well on your way. Jump back in whenever" (plus a few emojis for fun). And they did! They came back and finished the rest of their onboarding plan with the help of the agent. They went live. That manual hack is now the next thing we're building into the agent. Here's the thing I keep thinking about: in 15 years of shipping software, I don't think I've felt this kind of excitement about a launch before. Not because it was a big launch. But because I genuinely didn't know if it would work. In hindsight, I was getting bored being a PM. With pre-AI software, once you knew what to build, you knew it would "work". It was just a matter of time and effort. Now with AI, you can have good instincts, watch the thing run, think "yeah, that feels right" and still have absolutely no idea if a real human will actually engage with it. That level of uncertainty used to be the thing I wanted to eliminate. Now I think it's the thing that makes this era of building so alive/anxiety inducing/fun. I'm going to keep writing about this: what it actually feels like to build AI products from the inside, mistakes and scrappy workarounds included. None of us have done this before. We're all figuring it out. Might as well do it out loud.
-
Christine Zdelar shared thisMy first post as a Fin employee! Naturally, it's another major product launch. Tonight in San Francisco you can say hey to Operator, a completely new way to interact with our Customer Agent. Or, if you're like me, register at the link in comments and watch the recording over tomorrow morning's coffee
-
Christine Zdelar shared thisFeels like I was just on here telling you about a new role for Fin, Intercom's Customer Agent. Well, it's a new week, so why not a major new Fin launch? Introducing Fin for Ecommerce From the first browse through every repeat purchase, Fin supports the full ecommerce journey. It helps customers in the moment, answers questions, suggests the most relevant products, and smoothly guides them to checkout. It works like a top-tier store assistant with deep product knowledge. Delivering at scale, right out of the box. Hear directly from Kurt D. at Avocado Green Brands who entrusts Fin with, in my opinion, one of the most important things you can spend your money on (a really good mattress ) What should Fin's next role be?
-
Christine Zdelar shared thisWelcome to the world, Fin for sales. As part of our vision to build a unified Customer Agent, Fin now acts like your best inbound sales person. We've brought Fin to the start of the customer journey: qualifying leads, booking meetings, closing deals. This is a huge day for Fin, but you have to know, there's so much more to come. More roles. More outcomes. More Fin. And I personally can't wait to show you what's up next for our Customer Agent...
-
Christine Zdelar posted thisIt's crazy to think that a year ago this week we started building out Insights at Intercom, my first foray into deep AI product building. (And crazier to think that now this week I'm also starting on a new venture for Fin... watch this space). But reflecting upon the last year with Insights all I can think about is scale, scale, scale. AI can solve problems to a degree previously unthinkable. While I *knew* this before, it wasn't until we started building Insights that I realized what this new unlock meant for me, the PM. Now, my job is all about confidence building You can read more about what I mean at ideas dot fin dot ai. Link in comments.
-
Christine Zdelar posted thisIntercom is coming to Berlin .... next week! On Wednesday, December 10th come meet me and a few other members of the Intercom team to learn all about building Fin and what it it's like to work at Intercom. What's on offer? - Lessons learned building AI products - Live demos of what we've shipped this year - an AMA about what it's like to work at Intercom R&D - most importantly, free coffee Register to come meet us at the link in comments
No more next content
Slide to item 1
Slide to item 2
Slide to item 3
Slide to item 4
Slide to item 5
Slide to item 6
Slide to item 7
Slide to item 8
Slide to item 9
No more previous content
-
Christine Zdelar liked thisChristine Zdelar liked thisDo you have clear and strong beliefs about the future? Does your leadership team? It is impossible to create an excellent product strategy without clear beliefs about the future. And if you do have clear beliefs about the future, everything gets easier and better! Strategy becomes clear, employees get excited, execution gets faster. This is more important now than any time in the past 15 years because the future is so open and it is easy to get lost. Companies without clear beliefs about the future will fail. They will be unfocused internally and incoherent externally. The company will argue endlessly about what to do, and execute slowly as people question what they are doing, and why they are doing it. Sounds obvious, yet companies with clear and strong beliefs about the future are a minority. This is because it is hard. No one knows what the future holds, and people are very bad at predicting it. That's why every vision of the future had flying cars, and I'm still looking up and can't see any flying cars. If you have strong beliefs about the future, you can build products against those beliefs, and turn that into differentiation in the market. Building strong beliefs requires three things: 1. Research and synthesis. Leaders must consume and learn in great volume: Read and understand the frontier of technology. Use products at the cutting edge. Understand where your customers are at, what they believe. Find patterns. 2. Conviction. Build strength of conviction through heavy debate with others who also research heavily. Debate and converge. Spend time doing this, almost every day. 3 Fortitude. No matter the research and conviction, you know you will be wrong about some things. But you have the strength to pursue your dreams and goals nonetheless. There is an important thing here: you need both open mindedness about being wrong, and the fortitude to continue pursuing executing against a belief, even in the face of many saying you're wrong. Most people will be wrong too, many will be wrong about you being wrong! Concrete example: Fin. We have five very clear beliefs about the future in our category, and we build against them as our strategy. No surprise that those five beliefs are our biggest product differentiators. I'll share a short version of each one in a comment. If you're curious, please ask questions on any of this, how we approach product strategy, our own individual beliefs and how they came to be etc. I'd love to help others do this well, these can be challenging times!
-
Christine Zdelar liked thisChristine Zdelar liked thisYour AEs and SDRs are expensive. How much of their week goes to leads that were never going to buy? In most evaluations I run, it's the majority. Reps working through inbound that will never convert, while the leads worth having a conversation with sit in a queue behind them. It doesn't show up in any funnel report, it shows up as flat pipeline. We had the same problem at Fin. 10 SDRs, 24/5 shifts, manual triage on everything. Since November, with Fin for Sales: - 40% of the SDR role now handled automatically - Rep conversation load down ~50%, from 1,275 a week to 675 - MQL first response SLA up from ~75% to 95%, no extra headcount - $3M influenced closed-won, +44% SQP, +95% YoY pipeline ARR The same 10 SDRs stopped triaging and started multithreading, re-engaging cold prospects and working closed-lost. Full story in the comments :)
-
Christine Zdelar liked thisChristine Zdelar liked thisWe've just announced our Evals and Releases product. The AI agent product space has matured, and customer service teams now need to behave a lot more like our own AI engineers in how they test and release improvements to Fin. Evals and Releases give you: - full e2e testing capabilities across your entire Fin configuration, using LLM as a judge for both subjective quality and objective pass/fail criteria - complete safe, insulated test environments - a/b testing and controlled rollout These are sophisticated features, but as always, we really sweat how we build them so it's easy for your team to use. And Operator is baked in from day 1. Which makes it even easier. Bottom line: you can now make changes to Fin much more safely and much more confidently. This is the way to high automation rates. As always -- here's a demo, so you don't have to take my word for it!
-
Christine Zdelar reacted on thisChristine Zdelar reacted on thisZendesk just added two more headstones to the graveyard. This month's funeral notices: AI agents Essential and the legacy bot builder. Development stops 31 August 2026. Full shut-off 10 December 2026. That's on top of legacy custom objects (buried 27 July) and a batch of Built-by-Zendesk apps due 21 October. So if your support automation lives on any of those, you're rebuilding it. That decision's been made for you. The only decision left is WHERE. Here's the thing nobody at the funeral wants to say out loud: you're being asked to rebuild your bots on the same platform that just buried them. Same effort, same flows, same graveyard - just a fresher plot. Or you rebuild once, somewhere that won't sunset under you: Rebuild in Zendesk: same flows, new tool, wait for the next headstone Rebuild in Fin: AI-native, no legacy layer to bury - the agent IS the product Not a like-for-like copy - Fin resolves tickets end to end, not just deflects them Voice, multilingual, and your historical data brought across, not left behind Rebuilding your automations is the expensive part either way. The 10x is choosing what you rebuild them INTO. We move teams off Zendesk for a living - full history, contacts, tickets, the lot (Sabel + Migr8Now). If December's deadline is already on your calendar, let's chat before you dig the next plot. #Zendesk #FinAI #CustomerExperience #SupportOps
-
Christine Zdelar reacted on thisChristine Zdelar reacted on thisIt feels strange writing this, but today marked my last day at Fin (formerly Intercom). Looking back, it seems fitting to share this photo from my very first day. Just a week before starting my first big job as a Customer Support Engineer, I broke my ankle playing rugby. Walking into the office on crutches wasnt quite how I imagined beginning my career, but at least it makes for a pretty memorable start and also worked as a fantastic ice breaker Over the past almost three years, so much has changed. We moved offices, I grew from Customer Support into a Product Support Engineer role, and I had the chance to learn from some incredibly talented people. More importantly, I made lifelong friends and shared countless laughs. A huge thank you to everyone I had the pleasure of working with, to every manager who supported and believed in me along the way. Youve all played a part in shaping my first chapter in tech and Im incredibly grateful for it. Now its time for the next adventure. Im moving to Australia with no job lined up, but with plenty of excitement for whats ahead. Its a little scary, but thats part of the fun. Thank you, Fin, for a truly special first chapter. Ill be taking all my memories with me for a long time to come. Heres to new beginnings
-
Christine Zdelar reacted on thisChristine Zdelar reacted on thisWeek 1 in Fin behind me Excited to be back to the company that I know well, and at the same time has evolved in so many ways to win the shift to AI! Super grateful for opportunity and confidence from Darragh C.rom Darragh C., Paul Phillips and Jordan Neill, and pumped to create the best place to build AI products in Berlin. Fin has a long track record of creating exceptional teams, and sharing our knowledge and experience with broader community. We will do the same in Berlin We have a lovely office in the Mitte. If you are local in Berlin and would like to talk AI or product engineering, host small meetups, get a speaker or two for your meetup - give me a sign!
-
Christine Zdelar reacted on thisChristine Zdelar reacted on thisOur new product Operator goes live to all 12,000 Fin customers today. The feedback has been exceptional. Its one thing to make AI work for a handful of customers via consulting/pro-serve, but deploying as a self serve product to 12K people is a different challenge. Operator is (imho) prototypical of where software is going, in 4 specific ways 1. It removes a huge % of the tool time (the pointy/clicky/draggy stuff that most products need, Operator just does it) 2. It renders UI dynamically to solve problems and visualise data as you need it. 3. It expands the potential user base of the product. You already know how to use Operator, there is no learning. 4. It elevates software from task based work to strategy based work, which in turn elevates its users. Link in the comments.
-
Christine Zdelar reacted on thisChristine Zdelar reacted on thisOperator is now available to every Fin customer! This has been an incredible product to work on - the demand and positive feedback have been unlike anything I've ever experienced. When you have customers tell you that "they're getting an Operator tattoo" you know you're on to something :)
-
Christine Zdelar reacted on thisChristine Zdelar reacted on thisOperator is now out of beta and available to all our customers. Customer response to Operator is unlike anything we've shipped since Fin. These are genuine comments from customers: "Using Operator feels like going to Disneyworld for the first time." "I'm getting an Operator tattoo." "This is the best product you have ever built. Operator is incredible." So, Fin is our AI Agent for your customers, answering their questions. Operator is our agent for your support operations team. It helps them understand and improve their entire support operation. Have a watch from these customers describing how they've used Operator (links to more short videos in the comments.) If you're a customer, get stuck in now and see how Operator will dramatically accelerate your support ops team. And get the full scoop at fin.ai/operator
No more next content
Slide to item 1
Slide to item 2
Experience & Education
-
Fin
******** ** ******* **********
-
********
******* ****
-
*******
**** ******* **********
-
******* ******* ******
****** ** ******* * *** ***** ********* ********** undefined
2019 - 2020
-
*** ******* ** ******* *** ****
******** ** **** ****** ****** ******
2008 - 2012
View Christines full experience
See their title, tenure and more.
Sign in
Welcome back
By clicking Continue to join or sign in, you agree to LinkedIns User Agreement, Privacy Policy, and Cookie Policy.
New to LinkedIn? Join now
or
By clicking Continue to join or sign in, you agree to LinkedIns User Agreement, Privacy Policy, and Cookie Policy.
Volunteer Experience
Publications
-
How has Covid-19 and remote working changed millennials?
Business Graduates Association Sep 2020
-
How Millennials Made Sense of Transactional Distancing to Maintain Work Performance: A COVID-19 Investigation
Academy of Management, Organisational Behaivour Division Aug 2020
Recommendations received
-
LinkedIn User
“Christine was my director on an emerging product team when we were both at Politico. I was instantly and consistently impressed by her ability to navigate complex interpersonal challenges, all while instituting and promoting product management best practices. Christine demonstrated expertise and strategic know-how in introducing various internal stakeholders to key product management concepts, often in circumstances where others had failed before her. Under her leadership, our product team saw incredible improvements in key metrics, and gained the respect and appreciation of other company leaders along the way. Her ability to remain flexible where necessary while also pushing for important changes was a huge benefit. I am grateful to have work with Christine already, and would be excited to do so again. Any company looking for a dynamic leader with product management expertise and a mastery of people management need look no further. ”
3 people have recommended Christine
Join now to viewView Christines full profile
-
See who you know in common
-
Get introduced
-
Contact Christine directly