Apr 14, 2026, 8:00 AM ETInterview

From SEO to Agent-Led Growth: Profound's James Cadwallader

Marketing has shifted from a "nice-to-have" to an absolute necessity; teams not utilizing AI agents, specifically "profound agents," for execution are failing.
Profound serves approximately 12% of the Fortune 500, helping companies navigate visibility across AI search models like ChatGPT, Claude, and Gemini.
The market is undergoing a fundamental platform shift from "blue link search" (deterministic human discovery) to "probabilistic AI responses" (agent-mediated discovery).
The primary change is not the internet's interface but the "user": agents now crawl the web to discover, verify, and recommend brands before humans ever see them.

Agents consume content differently than humans, utilizing the long tail of the internet; James cited a personal example where ChatGPT used 65 different web pages to answer a simple query about a showerhead.
Future AI agents are expected to use orders of magnitude more of the internet to synthesize answers, driven by Moore's Law and decreasing inference costs.
Content strategy must shift from building for human attention (SEO) to building for agent discovery and consumption; much of this new content may never be read by a human.
High-consideration categories (consumer electronics, autos, white goods) are seeing the most immediate impact from agentic search due to the complexity of research required.
Platform-specific behaviors vary significantly based on training data:
- Gemini: Leans heavily on YouTube content due to Google ownership.
- ChatGPT: Frequently cites Reddit for consumer queries and LinkedIn for B2B.
- Claude: Historically relied on pre-trained models but has recently updated classifiers to prioritize real-time web data more heavily.

James warns of a potential "dead internet" scenario within three years, where the economic incentive to publish content diminishes as human traffic disappears and agents vacuum information without clicking citations.
The current advertising model relies on human page visits for revenue; if agents bypass these pages, the economics of the open web could collapse.
Potential solutions include vertical integration of AI labs with social networks (e.g., OpenAI acquiring Reddit) to secure real-time human data, or platforms like Reddit and X monetizing agent scraping via API deals.
Social media is predicted to become the primary "biozone" for human status and idea exchange, while the broader internet remains an agent-accessible data layer.
Current attempts to "game" the system via self-serving listicles are effective for now, but models are expected to evolve toward first-principle reasoning, making such tactics less viable over time.

Generative advertising will likely become the most effective format, utilizing hyper-personalized context within conversational interfaces where consumer intent is highest.
Future ad units may function via system prompts where advertisers define target audiences and specific conversation triggers rather than traditional banner placements.
B2B and developer contexts differ significantly; agents are expected to remain objective and unswayed by branding, demanding technical interoperability and data legibility over emotional persuasion.
High-quality content is not a "red herring"; the key to winning in the agentic era is providing "original insight" that the models do not already possess in their training data.
Brands must treat agents as "super-intelligent beings" and ensure their products are legible, usable, and interoperable for automated troubleshooting and recommendation.

Visibility is the prerequisite for optimization; marketers must use specialized tools to analyze exactly how and where their brand is cited in AI responses.
The misconception that agents merely automate existing workflows is incorrect; they enable entirely new forms of marketing previously impossible, such as real-time battle card generation from sales transcripts.
James asserts that not using agents for content creation and distribution is functionally equivalent to ignoring the internet entirely.
A core strategy for 2024-2025 is to "equip super intelligence" with unique, proprietary data about new products to ensure accurate recommendations.
Marketers should stop trying to pattern-match to old SEO levers and instead focus on building content specifically designed to be consumed by algorithms.