I'm Yanina Ojani, and I'm really excited, to be here and to talk to you about AI, a great first date but a terrible marriage.We'll be about we'll talk about why pilot projects fail to scale, for enterprises. This session will draw on my twelve years experience advising Google's largest enterprise customers on how to build and how to execute transformational AI strategies. Over the next roughly fifteen minutes, because I hope we'll have, like, five minutes at the end to, for discussions and questions, I really hope to inspire you, and I hope to give you some very hands on practical tips on how you can turn your AI and AgenTek projects into an enterprise capability.

But before we dive in, let me tell you a little bit about myself, and let me double check that you can all hear me. But I think it looks pretty good. Awesome. Thank you, Gabriela. Yes. Please feel free to ask your questions in the chat. I will, answer them all at the end of the session. So I'm Yanni, and here are three stats about me. So seven represents the number of countries I lived in. I'm originally from Germany. I'm based in New York, and I lived across Europe, Africa, Brazil, China, and many more. And what does 96% means? 96 means that 96% of Google employees have started at Google after me. So I'm basically a Google dinosaur with my twelve years, of Google experience. And while I had many different roles, I had three major career pivots.

I've spent most of my Google career in the Silicon Valley, working with enterprise customers on big marketing or go to market transformation projects. But then with the explosion of Gen AI, I really wanted to be closer to the underlying tech infrastructure. So I pivoted to the Google Cloud side of the house, and I took over the role as a sales leader for the market introduction of Gemini. So it was just at the time when we moved from BART to Gemini, and I was really leading, that market introduction, for the Northeast Of The US. And now I'm leading the Google Cloud executive, partnership program, and I will draw a lot from my experience as the Gemini lead and my current role as well. Awesome. Let's dive in. Why are we here? What's the challenge? Let's talk a little bit more about the challenges enterprises are facing when it comes to AI.

Because do we not hear the great promise of AI? It will make all our lives so much easier. It will help us to reach new levels of productivity. It will help us to deliver new personalized and customized customer experience. And we can all do all of this by operating more cost efficiency. Isn't this great? Yet we see 80% of AI projects fail. That's a really big number. Just opening the chat here to have that on the site. So why why is that? This is a study from McKinsey which shows that 80% of AI projects fail to show any impact on the bottom line. And if you look around and you look at other studies, you see a lot of similar numbers. Long story short, spoiler alert, this is not due to the underlying technology. This is not due to the models. It's even not due to the costs. Right?

It's really due to the lack of change management. It's due to the lack of the required organizational structures, business processes, and the leadership which has come with it. So as the title says, I like to compare this to AI is a great first date, but it's a terrible marriage. Let's explore this a little bit more. When you go on a date, everything is really exciting because it's a very isolated experience. It's very low stake, and you can focus solely on the immediate chemistry. It's similar with AI. When you do a pilot, it's a very isolated environment. But when you move to a marriage or when you move to an enterprise solution, it requires so much more. Let's explore five key areas. Spark versus system. Right? It's pretty easy to get a first wow moment from AI. You put something in a chatbot and you're like, wow.

This just saved me ten, twenty, thirty minutes, an hour. But when you move towards an enterprise solution, you really have to integrate it across your entire enterprise, all your different enterprise tools. Right? And we know you have millions out of there. And this can be very often kind of messy, expensive, and complicated. Second part is lack of shared memories. A quick date, you don't need to remember your in laws birthday. You do have to do that when you're in a marriage similar to AI. Right? Like, when you do a pilot, you AI doesn't need the full historical context and the full customer experience. But when you scale AI, it does. And then there, you have the challenges that your data and your workflows are usually in silos, and it can create major hurdles for success and for integration.

Cost of living. Small pilots are cheap, but running AI enterprise wide brings a lot of costs with it if you look at ongoing cost of compute, maintenance cost, licenses cost, etcetera. And that often just requires a really deep financial analysis and an RI study, which many companies don't have yet. Safety and trust. You maybe start a pilot with non sensitive data or very kind of isolated small set of data. But a full, integration really requires you to open your book, and that triggers very often massive pushback from governance teams, privacy teams, or security teams. And last but not least, the underlying foundation change management. Long term adoption, really often fails because companies are just not ready to fundamentally change how they operate.

It's not about tweaking this or that. It's really about fundamentally changing how you're operating from the top down and from the bottom up. So don't get me wrong. Dating is great. You should run all these pilots. Right? You should get your feet wet. You should get all this first data. But now we're in 2026, and now we need to think, how do we integrate AI into a long term enterprise capability? So let me take a break. And usually, I talk about three pillars now. First one is a strategic executive alignment. That's really the organizational structure. How do you get your company AI ready? Second is designing agentic workflows. As I said at the beginning, we often start with individual, tasks in these wow moments. But if you really wanna, realize the full value of AI, you really need to look into your processes and your workflows and translate them into fully agentic workflows.

That's when the value is gonna be realized. And third, scaling it, securely. We won't have time to go through all of that. So I'm just gonna focus on the first one, which is the organizational structure, as that's kind of the first and most important piece of the puzzle. So when I started my Gemini sales role, I was really excited. Everyone wanted to talk to me. Everyone wanted to talk about the latest models and the latest tech, and I had so many meetings, and it was great. But then when I moved to procurement or deployment, it all looked very different. Right? Then the security team had said something to say. Finance team had something to say. All the different teams had suddenly, like, were involved, legal teams, etcetera, etcetera, and it was really difficult to move to to deployment.

So why are we talking so much about the models when we saw the stats at the beginning that the AI success really depends on, like, so many other factors? And that is the case because AI conversations are usually lacked by the IT team. Right? CTO, IT managers, etcetera. However, we really need to realize that AI is not an IT project. It's a fundamental shift, and that needs to be integrated throughout all layers of the organization. So it needs to be a priority not only for the IT team, but it needs to be a priority for the business leader, for the practitioners, for the c levels, and they all need to be part of the journey. So that's why I'm gonna go through the different personas and really elaborate what you can specifically do or that persona can do to advance your company on that AI journey. And only when all these personas play together, you will see long term AI success. So let's start with the CXO. So the CXO is really the strategic architect. Right?

They are setting the vision. Their task is to remove friction in the organization and really lead by example. So the first, focus points for them is AI business alignment. So they clearly outline the top priorities for your company, and then their goal is to identify exactly where I could act as a multiplier for those goals. Second one, really important, is to establish this cross functional AI center of excellence. So many companies adopt a center of excellence nowadays, but the crucial part that it's cross functionally. Right? You wanna mandate that legal, compliance, HR, and all your other, parts of your organization will be included in the center of day one. That really avoids the wall of no, which I mentioned earlier, and really helps your speediness, to deployment and really helps you to look holistically, at AI across the organization.

Number three, also one of the most important, art parts is to really modernize your data foundation. Right? You need to shift from viewing data as like a byproduct to treating it as a very strategic asset since that's kind of at the heart of AI, and it will really help you to create a unified accessible strategy across your entire business. We need to lead the cultural shift, foster psychological safety. People will not get it right from day one. They need to know that it's okay to have, like, little micro failures to really, realize the full value and potential of AI. And last but not least, you need to operational AI. You need to lead by example. You need to share your use cases in all hands meeting. Have your EAs or ABPs, share share their use cases. AI is a great tool for business partners.

And really signal to your organization that AI is a tool for everyone and not just for IT. Let's move to the functional leaders, head of. They are really the operational translator. Right? They are there to scale use cases and build that connective tissue between different teams and between practitioners and the big vision from the CXO. One is empower AI champions. So we often recommend to build your little AI champions network, have set communication channels at that Slack or Teams on Google Meet, whatever you use, for them to really amplify the message, for you. You will automatically have people in your team who love to research about AI. Let them share their knowledge and really create a train the trainer model. Bridge the skill gap to the analysis of the AI skills within your team and really identify functional gaps and provide targeted upscaling. So you're not leaving anyone behind.

Incentivize best practice sharing. Right? Maybe, create a prompt library or some kind of resources where people can share their their learnings. And then also incorporate regular reviews of AI initiatives. Really see on, like, what's happening in your team and which are the initiatives which can really bring, long term AI value and add more resources behind them. Last, personally, the practitioner, the individual contributor. They are really eye professionals. They know their work and their workflows the best, so they need to develop use cases. They need to, increase a personal productivity. And then very important, they need to be responsible when they execute. So they need to master the prompt literacy. I talked about the prompt library before.

They need to commit to, like, micro experimentation, make it part of the OKRs, make it part of the one to ones so they're really committed to test and learn. And, ultimately, they want you want to help them to design AI native workflows. Right? Stop treating SI as kind of the final check. I just gonna check that output at the end. I kind of check for mistakes or whatsoever, but really encourage your teams to start using AI on day one, really as a collaborator, as a brainstormer, and help them put something to paper. Test and iterate. And then very important, last but not least, really adopt that human in the loop guardrail. Right? Ensure that everyone still feels responsible for the outcome and that they verify the outcome for bias and for accuracy, that they don't put PII data into non enterprise solutions, etcetera, etcetera.

Once you have all of that set up, your organization really well set up for long term AI success and really, realize the the value of AI. So AI strategies are not built on, like, a spark, but it's really the shared commitment between your entire organization, which will help you to be long term successful. And with that, I'll