From Keywords to Prompts: How Otterly AI Is Building the Brand Visibility Layer for AI Search

In this week’s episode of Voices of Search, we spoke with Thomas Pam, CEO and co-founder of Otterly AI, about what it actually takes to monitor and optimize brand visibility in a world where 93% of AI Mode interactions result in zero clicks and 58% of Google searches never leave the results page.

Thomas spent nearly two decades in marketing—including leading a 30-person team at headless CMS company Storyblok—before building Otterly out of a simple frustration: no tool showed him how his brand was showing up in ChatGPT. Today, Otterly has over 20,000 users, a Gartner recognition, and a growing ecosystem of partnerships with Semrush, Press Ranger, and Storyblok. And they’ve done it entirely bootstrapped.

Thomas explained why AI search is a genuinely distinct category from SEO, how marketers need to rethink their KPIs from rankings to brand visibility, and why entity harmonization across the web may be the most underleveraged lever brands have right now.

Key Takeaways From This Episode:

• AI search is a distinct channel requiring distinct tooling—just as social media monitoring emerged as its own category, AI search monitoring is doing the same, and conflating it with traditional SEO is a strategic mistake.
• Brand visibility, not traffic or rankings, is the defining KPI in AI search—and that shift has significant implications for how SEO and brand/PR teams need to work together.
• Entity harmonization across the web (your CMS, press releases, community platforms, third-party listings) directly affects how confidently LLMs identify and represent your brand in responses.
• Bootstrapped focus is a feature, not a limitation—while VC-backed competitors build monolithic suites, Otterly is betting on a best-in-class observability platform that pipes insights into the tools marketers already use.
• Wikipedia remains one of the most underutilized signals for AI visibility—Otterly created its own page eight months ago and saw measurable impact.
• Brand mentions matter more than links in AI search—being cited as the answer when someone asks for “the best running shoes” beats being footnoted in a training plan article every time.

Why This Isn’t Just SEO With a New Name

Thomas was deliberate about framing Otterly as a new category, not a tab on your Semrush dashboard. His analogy is instructive: social media monitoring tools didn’t emerge as a feature inside Google Analytics. They emerged because social was a genuinely different channel with different mechanics, different signals, and different optimization levers. AI search is the same.

“Those search engines fundamentally work differently for us as consumers,” Thomas said. “The optimization techniques are SEO-influenced, but it’s not only about SEO.”

He also pushed back on the silo problem that has long plagued SEO tooling. Historically, SEO data lived in its own corner of the martech stack, disconnected from CRM, content, and PR systems. Thomas’s concern is that AI search tools are heading in the same direction—and he’s building against that pattern through active technology partnerships designed to pipe Otterly’s insights into the systems where marketing decisions actually get made.

That’s why Otterly’s first major partnership was with Semrush, followed by integrations with Press Ranger for PR distribution and Storyblok for content management. Each one reflects a belief that AI search performance is a cross-functional problem. It shows up in your content, your PR, your community presence, and your on-page signals simultaneously.

The Prompt Research Problem

One of the more underappreciated shifts in AI search is how fundamentally different prompts are from keywords. Marketers spent years training themselves to think in terms of search queries—short, structured, often stripped of context. 

Prompts are the opposite: conversational, goal-oriented, and highly variable by platform.

Otterly approaches this through what Thomas describes as a structured prompt research and ideation workflow—helping teams discover, categorize, and organize the prompts they should actually be tracking, by market, by product category, and by language. This is the first step in the platform’s user journey before any analytics layer is applied.

“Consumers use those platforms fundamentally differently,” Thomas said. “We need to think about prompts—and how we show up for those different prompts.”

From there, the analytics layer focuses on data quality above all else. Thomas argued that the proliferation of AI visibility trackers has created a noise problem: many tools offer some version of brand tracking, but the signal quality varies enormously. Otterly bets that marketers can only act on data they trust, and untrustworthy data leads to wrong optimization decisions.

Brand Visibility Is the New Primary KPI

Perhaps the most important strategic reframe in the conversation was around measurement. Thomas laid out a three-bucket model for thinking about AI search performance:

• The first bucket is brand visibility: How prominently and accurately your brand appears in AI responses across ChatGPT, Gemini, Perplexity, and others. This is the primary KPI most of Otterly’s customers track.
• The second bucket covers traffic signals: AI crawler engagement on your own properties, human referral traffic arriving from AI platforms, and indirect branded search traffic from people who encountered your brand in an AI response but didn’t click through immediately. Thomas noted that roughly 15% of Otterly’s own traffic over the prior two weeks came from AI agents—a figure that would have been close to zero 18 months ago.
• The third bucket is the long-term goal: revenue attribution. Tying AI-driven brand awareness back to CRM data and pipeline. Thomas acknowledged this is still an unsolved problem for most organizations, but predicted it would become table stakes within the next 12 months.

The broader strategic implication is one Thomas found genuinely exciting: AI search is pushing marketers toward brand in a way that performance-focused SEO never did. “SEO teams have historically optimized more for performance over brand,” he said. “We now see this big shift toward brand, which makes it exciting for brand marketing teams, but also means we need to cut down team silos.”

Entity Harmonization: The Most Overlooked Lever

If there was one concept in the conversation that felt underappreciated relative to its actual impact, it was entity harmonization. Thomas described a pattern he’s observed across many Otterly customers: inconsistent or contradictory descriptions of a brand across different web properties—CMS pages, press releases, community platforms, third-party listings—that create ambiguity for LLMs trying to understand who the brand actually is and what it does.

“If that’s not harmonized, if that’s not cohesive across the internet, then ChatGPT has a harder time recognizing who we really are and what we really do,” Thomas explained.

This has both defensive and offensive implications. On the defensive side, brands that don’t control their narrative across ecosystems can find AI models surfacing outdated, negative, or simply wrong characterizations. 

Thomas shared a concrete example: customers discovering that ChatGPT was citing specific Reddit threads containing negative product reviews—content that had previously been inconsequential in Google search but now carried real weight in AI responses.

On the offensive side, entity harmonization, when done correctly, helps LLMs build a confident, accurate representation of your brand that surfaces consistently. It’s the kind of work that sits at the intersection of SEO, PR, content, and community management—which is exactly why it doesn’t fit neatly into any single team’s mandate today.

Bootstrapped and Building an Ecosystem, Not a Suite

A recurring theme throughout the conversation was Thomas’s deliberate choice to remain bootstrapped—and what that constraint forces in terms of product focus. While well-funded competitors are building what Thomas called “monolithic marketing suites,” Otterly is making a different bet: be the best observability layer for AI search and integrate deeply with the tools marketers already rely on.

“I’m not a big believer in building out monolithic marketing suites,” Thomas said. “Marketers don’t need yet another content tool. They already have plenty.”

The product philosophy is integration over accumulation. Otterly’s partnership with Semrush means AI search insights can surface alongside organic performance data. Press Ranger enables marketers to act on PR signals that Otterly surfaces. Storyblok closes the loop between content performance in AI responses and the CMS where content actually lives.

Thomas was candid about what this approach requires: discipline about scope, a willingness to do fewer things better, and a belief that the ecosystem play is ultimately stronger than the all-in-one play. He also acknowledged that AI has changed the calculus for bootstrapped companies—making it possible to move faster with smaller teams in ways that weren’t feasible before.

The Lightning Round: What Actually Matters

When Jordan moved into rapid-fire questions, Thomas didn’t let up—two years of building inside this problem will do that to you. 

Here’s what Thomas revealed:

• The biggest myth about AI search visibility? That it’s not influenceable. “Most of them use web search-type features of various degrees. I’ve seen big success, be it through on-page or off-page. You can get your content cited within ChatGPT in a couple of days.” Thomas’s advice: run your own experiments rather than accepting the myth.
• What signal are marketers underutilizing? Wikipedia. Otterly didn’t have a Wikipedia page 12 months ago. They created one roughly eight months ago and saw measurable improvement in how they appeared in AI responses. It remains a “hard nut to crack” but one with real payoff.
• Brand mentions or links? Mentions, every time. Thomas’s example: a running shoe brand could get their website cited in a marathon training article, or they could get mentioned directly when someone asks for the best running shoes. Only one of those has a direct business impact.
• If your SEO dashboard vanished tomorrow? Thomas said he’d be inclined to rebuild nothing—a deliberately provocative answer. The one metric he’d preserve: impression share on Google, but reframed as a brand-building signal rather than a performance metric. He doesn’t expect lost organic traffic to come back and isn’t building a strategy around the assumption that it will.

The Bigger Picture

What makes the Otterly story interesting isn’t just the product. It’s the timing and the discipline. Thomas launched in a category that didn’t fully exist yet, built to 20,000 users without external funding, earned Gartner recognition through product obsession rather than marketing spend, and made deliberate choices about what not to build.

The fundamental insight driving all of it is one that every marketer will eventually have to reckon with: when 93% of AI interactions end without a single click, optimizing for rankings and traffic is measuring the wrong thing entirely. 

Brand visibility—how you show up in the AI responses people actually read—is the new terrain. The question is whether your measurement infrastructure can see it.

Voices of Search is a daily SEO and content marketing podcast hosted by Jordan Keone and Tyson Stockton. The show delivers actionable strategies and data-driven insights to help marketers navigate the ever-evolving world of search engine optimization and content marketing. New episodes air weekly, covering everything from technical SEO to AI discovery, featuring industry leaders and practitioners sharing real-world frameworks and proven tactics.

Subscribe to Voices of Search on Apple Podcasts, Spotify, or your favorite podcast platform. Follow Previsible on LinkedIn for updates and subscribe to the VOS YouTube channel for video episodes and clips. You can also visit the official VOS site to explore the full episode archive and submit your SEO questions for future episodes.