What is generative engine optimization, and how does a business show up in AI answers?

Welcome back to another episode of GO AI. Our guest today is Leah Nurik, founder and CEO of Brandi AI. Leah, would you mind telling our audience who you are and a bit more about yourself?

Sure, thanks for having me today, Connor. My name is Leah Nurik. I'm CEO of a company called Brandi AI, which is pioneering generative engine optimization and how marketers are using AI in this new AI-first marketing era. Before starting Brandi, I ran a public relations and strategic marketing agency for about 14 years, and before that I led go-to-market strategy for Motorola's enterprise B2B application markets. I've been in software my entire career, helping to solve technology challenges in businesses, so this is a natural evolution. Some of the biggest tech problems for businesses right now have to do with AI, and specifically how AI can lead to sales for the business, which is where GEO, generative engine optimization, comes in.

Before we get into the weeds of GEO, I want to clarify for our audience. There are many acronyms that have popped up: AEO, AIO, GEO. How do you define GEO?

Generative engine optimization is how you optimize your marketing, your answers, your website and everything you produce for generative engines. Think about how buyer discovery has shifted in just the past two years. You used to get a blue link when you searched for products, whether B2B or B2C. Now you get an answer returned, whether you're using the traditional Google search bar or going to ChatGPT, Gemini or Grok. Generative engine optimization is how I optimize my approach to marketing, content and messaging, across all the disciplines in the marketing organization, to make sure I'm front and center in buyer discovery, and also that I'm positioned correctly. Before, a blue link was a blue link. It was either the right blue link or not. Now it's an answer packed with sentiment, accuracy, potential misinformation and storytelling. That's a very different practice, and a brand-new one marketers have to deal with because of synthesized answers.

It's a bit like SEO in that there are a lot of influences on GEO: do this and you better your odds of ranking well in a generative engine, do that and you lower them. What are some of the well-known influences business owners should know about, and what are some lesser-known ones?

I would differ with the SEO comment first. SEO was more about technical structure and backlinks that formed authority. That is very different now. You do need to structure your data to be found, but it goes further: once the AI finds it, how does it read it? Is it reading the information from multiple sources?

As for the unexpected things, one is public relations. Traditionally, PR didn't really affect SEO. Maybe a backlink from an article helped your authority. Now PR and earned media play an incredibly important role in most markets. It's important to understand your market. Whether you're building B2B software, making running shoes, selling candles or running a landscaping company in North Carolina, what the AI finds authoritative, and what it uses to return your brand, is unique to that market. So seeing which domains are being cited is number one. We can see that more than 80% of the domains that get cited and influence the answer come from unpaid sources. That means authentic, credible content production and publishing, and customer review sites from Trustpilot to G2, play a role. Authentic storytelling in multiple places, and truly delivering on your mission so people want to leave good reviews and talk about you in different corners of the internet, is what helps return you in AI search results.

There are also multiple generative engines. In the days of SEO you could say you were mostly focused on Google. Now we have Perplexity, Gemini, ChatGPT, Claude and Google's own AI Overviews. Are you finding in your analytics that a practice works well for Perplexity but not for ChatGPT, or are the results similar across engines?

The structure for reading is similar. The AI reads a page in chunks and looks for answers to questions, and that's pretty consistent across the board. But the domains the different platforms cite to answer a question can be very different. On some models Reddit is still super important, and on others Reddit is down in the top 40 domains instead of the top 10. Same thing with YouTube, and you can probably imagine which platform YouTube performs well on. LinkedIn tends to do really well on Copilot. So it's about looking at where the AI is getting the answer from, then building a cohesive strategy that targets those locations.

Do you vary GEO strategy by industry? A car dealership versus an insurance company, for example?

The citation sources for each market can be vastly different. If you're in restaurants, Yelp might come up a lot. It won't come up if you sell B2B software. If you sell health supplements or makeup, public Facebook groups or Instagram influencer reels might come up. Every industry is unique, but here's one interesting example. Some industries don't traditionally focus on their own content. High-end fashion brands and car brands tend to have brochure-style websites. We've run a demo instance for Jeep for over a year, and when I look at its cited sources, the AI doesn't pull much from the brand and competitor websites, because those companies don't produce a lot of owned content. More than 90% of the cited domains are earned media and third-party review sites. Is that because the AI finds earned media more credible, or because brochure sites aren't structured for readability and don't produce much content? It's a chicken-and-egg question. What matters for the person trying to improve visibility is figuring out where the AI is pulling from, how to maximize that, and whether to rethink the approach and start producing owned content across different channels. When you have the insight and the data, you can take action that makes a fairly rapid impact.

In SEO, a Google algorithm update could throw practices off balance and you had to adapt. Does GEO shift under your feet with new model updates, or is it mostly consistent?

With SEO, a lot of it was tactical, almost like gaming the algorithm. GEO is different. If you try to trick the AI, a model update can leave you stranded. Say you were paying people under the table to post on Reddit, and then a model deprioritizes Reddit. What we tell our customers is that it's not about gaming it. It's about telling a cohesive, authentic story across all these channels. The purpose isn't only to increase how often you come back, it's to make sure your story is in all the places that will future-proof your brand. Tell authentic stories in multiple ways, ask your customers to post reviews, and bring products to market that fulfill the brand promise. There are still insights and actions, like whether to hire a PR firm to amplify your story, but that's very different from pressing a button to trick it.

The models do change. We have full-time data scientists on staff and our own proprietary IP that crunches the compounding data and responses across our ecosystem to identify what's changing, what the models reward and what they don't, and we keep our customers and the product updated. It's about a future-proof brand strategy and fulfilling the mission you're telling the world. There are no tricks. It's basically what your mother told you in kindergarten: tell the true story and deliver great relationships.

What have been your biggest discoveries since founding Brandi AI?

One thing I find really cool about GEO is that the brands that get rewarded are the ones that speak with authority. With SEO, you were rewarded for authority when people linked to you. What we see now is that when you publish something that isn't fluffy, that's truly authoritative and helpful and answers a question in a way that isn't self-serving, you build authority and you come back more often. Understanding which brands and domains the AI considers authoritative, and making sure you speak to them, helps increase that. When you're mission-driven and telling an authentic, credible story meant to be helpful, you're rewarded.

The other surprising thing, which is very good for marketing in general: a lot of communications and content people thought AI would put them out of work. What we see is that AI doesn't reward AI-generated content at scale. Unique points of view and well-crafted, human-driven stories with empathy and context about the problems you're solving perform best. And earned media, third-party objective sites and research sites are acting as sources of truth that drive the answers. With SEO there were a lot of tricks. That doesn't seem to be the case with GEO, and I think that builds trust in the answers we get. We can all use a little more authenticity and trust.

Are there specific KPIs you focus on in GEO, like number of citations or frequency of mentions across generative engines?

Both mentions and citations are important. We look at how often a brand comes back in answers relevant to its market and for different personas. Everybody wants to be mentioned in the answer. Citations matter because there's subtext and nuance in answers. When your content or your PR coverage is being cited, you're helping define the context of that answer.

We also look at share of voice. Traditionally, marketers calculated share of voice against their ten competitors, adding up to 100. Now there can be brands tangential to your market, or influencers, driving leadership in the conversation, and they could be potential partners. Then sentiment and accuracy. We have a patent filed on our sentiment hub, which lets you define what you want to measure against and see why you're being ranked positively or negatively, instead of a random score like 7.8. That can also work as customer research, instead of commissioning an expensive report that takes nine months.

What's a good number? That depends on the market and how you benchmark against competitors. In a crowded space like coffee, 20% share of voice means you're really doing well. 85% probably isn't possible. For a marketing agency, the percentages might be lower. The key to GEO is insight into the data, which then informs how you act.

What areas of GEO are not well understood yet?

The engines don't tell you what the questions are, the prompts. They've started sharing some sample prompts in webmaster tools, but those aren't necessarily indicative of actual prompts. Another open question is how advertising will shift. OpenAI is rolling out ads in ChatGPT answers, but they're clearly labeled ads and aren't affecting the answer itself. These are closed systems, and we'll know what the models decide to release. "What is the best coffee place near me" can be asked a million ways. Using signals and intelligence about real conversations, what people complain about, what excites them, what causes them to buy, might eventually give glimpses into what people are asking. There are also vendors saying they know what the prompts are, and there's no way they know, because nobody publishes them. There's a good amount of snake oil being peddled because people are scared their traffic has dropped. It's important to find agencies, service providers or software providers you can trust to tell you what's happening, and also what they don't know. That transparency matters.

Is there anything you wish more business owners knew about GEO before you begin discussions with them?

A lot of people are waiting to see how this shakes out and letting others pioneer, and they're losing a big opportunity. There are ways to experiment and try things, and not every company is doing it, so there's a real opportunity to gain a competitive advantage in this early market. If you want to future-proof your brand, crafting the narrative now can build authority for the future. Sitting back and waiting for other people to define it is a major risk. It would be better for smaller or more cautious businesses to take at least a small leap, whether they want to keep growing or get an immediate competitive advantage for revenue growth.

Good words of wisdom to leave us on. Thanks so much, Leah, for coming on GO AI. If listeners want to reach you, how do they find you?

You can find me on LinkedIn at Leah Gabriel Nurik, and you can find Brandi at mybrandi.ai.

Awesome. We'll have the links to Brandi AI and your LinkedIn in the show notes for this episode. Thanks again for coming on GO AI, Leah.

Thanks. Have a great day.