Your Brand’s New Boss? A Chatbot. | Brafton

There are only so many times a writer can pull off Skynet and HAL 9000 jokes without coming off like a luddite. Now, we’re not necessarily at Space Odyssey levels of AI taking over, but if you’re running a business, you need to get comfortable with algorithms making decisions for your customers. And it doesn’t care one bit about your latest quarterly report. 

A new class of gatekeeper is now running the show, and if you’re not measuring your influence with it, I’m afraid you’re already falling behind.

From a Shout to a Search to a Single Answer

For decades, we had Share of Voice. It was simple, brutish and effective. You just out-spent the other guy on advertising. Whoever had the biggest pile of cash to set on fire usually won the most attention. It was the marketing equivalent of a shouting match in a crowded room.

Then, the internet got sophisticated, and we graduated to Share of Search. This was a genuine leap forward. Yelling the loudest didn’t work anymore. So now, it was all about actively being sought out. This metric gave us a direct line into consumer intent, showing us which brands were capturing genuine curiosity and consideration. It was a powerful proxy for market share and a darling of boardrooms everywhere.

But the ground is shifting again, and this time the change is seismic. Generative AI platforms like ChatGPT, Gemini and Copilot don’t give users a long list of blue links to choose from. They deliver a single, synthesized, authoritative-sounding answer. There is no page two. You either make it into the model’s response, or you are functionally invisible.

What the Heck Is ‘Share of Model’? (And Why It Should Excite You)

Share of Model (SoM) is the new metric for this new era. It measures your brand’s visibility, prominence and sentiment within the artificial intelligence models that are rapidly becoming the primary discovery tool for consumers. 

But since those models have more of a mind of their own (albeit in a black box), our “shares” can account for consumer sentiment, just like before, but also for the model’s notion of your brand. 

If someone asks their chatbot of choice about sneakers and your shoe brand’s name doesn’t come up among the examples, you might need to rework your strategy. Why, you ask?

Think of it this way: LLMs are constantly Hoovering up the internet — news articles, reviews, forum discussions, your website, your competitors’ websites — to build their worldview. Your Share of Model is the result of that education. 

If the model sees you as a leader, you’ll be recommended. If it’s learned that your customer service is terrible or your products are outdated, that’s exactly what it will tell the user who asks.

And people are asking. A recent Capgemini study found that over half (58%) of consumers have already replaced traditional search engines with Gen AI tools for product recommendations. 

That figure has more than doubled from 25% in 2023. So, if you were telling yourself this fluke will pass, I’m sorry. This isn’t a future trend or a mood shift anymore. It’s a present-day tidal wave, and it’s changing the very nature of brand discovery.

How To Measure Your Brand’s AI Reputation

This all sounds a bit abstract, right? How do you measure what a machine “thinks”? I already feel weird enough asking ChatGPT, “What do you think of this?” Now, we’re supposed to track that?

Well, in a sense, yes. The term Share of Model is both a generic descriptor for this new metric and the name of a specific platform from the digital marketing agency Jellyfish that’s designed to track it (smart move, right?).

The Share of Model™ Platform is a first-to-market solution that essentially puts AI models on the psychiatrist’s couch. It systematically queries the major LLMs at scale to analyze three critical things:

1. Mention rate: How often does your brand even show up compared to competitors? How do those mentions differ across different regional markets?
2. Human-AI awareness gap: How does your brand awareness with real people compare to your visibility inside the models?
3. Brand sentiment: How are the models talking about you? Are you the hero, the villain or an irrelevant side character?

If you’re the villain, you may hope that Jellyfish is just some flash-in-the-pan startup. Sorry again. It’s part of The Brandtech Group, which is all-in on Gen AI. The group also includes Pencil, an end-to-end generative AI marketing platform that has already created over 1 million Gen AI ads, learning from a staggering $1 billion in media spend.

We aren’t talking about developers watching the revolution from the sidelines. These folks are building tomorrow’s advertising world.

But they’re not the only ones capturing this new metric. Plenty of other platforms are experimenting in this space.
Ahrefs recently launched Brand Radar, designed to track brand presence in the chatbots your audience is likely already familiar with, including ChatGPT, Gemini, Perplexity and more. It includes this nifty little AI citations dashboard from the Site Explorer section:

You can click into this big blue numbers to navigate to the Brand Radar section where you can explore deeper into these citations:

Not one to be left out of the SEO sea change, Semrush launched its own AI Toolkit, which focuses largely on these same chatbots. This tool goes beyond showing you who’s saying what, but also provides recommendations to help you address user needs, whether through your marketing strategy, products or overall business strategy. 

Winning the AI Popularity Contest

OK, so you can measure your Share of Model. What then? 

The insights you gain become the bedrock of a new, smarter marketing strategy that treats the LLM as a key audience in its own right. Sounds strange, I know, but hear me out. 

The goal is to “teach” the models to perceive your brand favorably.

By understanding how LLMs view their clients, Jellyfish and similar platforms have been able to reverse-engineer success. They’ve found that strategically introducing Broad Match keywords — informed by the language the AI models themselves use — and proactively adjusting marketing budgets can generate incredible results. We’re talking up to 380% more conversions at CPAs that are up to 55% lower.

So, we’ll need to drop some of our beliefs about paid search optimization. 

The insights from an SoM analysis can and should inform everything: the copy on your website, the alt-text on your images, your PR outreach and your content strategy. If the models have a misconception about your brand based on old forum posts, you need to flood the internet with new, authoritative content that corrects the record. From now on, you’re actively curating the AI’s curriculum.

You may not like or use AI yourself, but the fact of the matter is that no enterprise can afford to ignore a giant factoid and mood generator with 700 million users — and that’s weekly.

The New Rules of Visibility

The shift from Share of Voice to Share of Search was no doubt a big one, but the leap to Share of Model is a different beast entirely. The recipes that used to work don’t apply anymore.

We’ve moved from a battle of budgets and keywords to a battle of perception and narrative, where the judge and jury is a complex algorithm. 

Your brand’s entire digital footprint is now its resume, and AI is the hiring manager. The robots are already forming their opinions about you and sharing them with your customers. Isn’t it time you found out what they’re saying?

Note: This article was originally published on contentmarketing.ai.