Scrunch AI partnered with the team at Bain & Company on the research brief, "Marketing's New Middleman: AI Agents" (published 2025). The findings confirm what many marketing and digital leaders are already seeing: AI agents — including ChatGPT (OpenAI), Perplexity AI, Google Gemini, and Anthropic's Claude — are now a major gateway to brand discovery and purchase decisions, often replacing traditional Google Search and compressing the path to purchase. Bain's research found that AI-driven "zero-click" journeys are reshaping how consumers evaluate brands, with AI agents synthesizing answers from across the web rather than directing users to individual sites.
Below are the four biggest takeaways from the Bain brief, plus practical steps, tools to use, and how Scrunch helps you operationalize them.
Customers are increasingly turning to AI-powered agents to research options and get personalized recommendations — bypassing Google results, brand websites, and even sales reps. These systems don't just read your site; they synthesize answers from third-party sources, reviews, earned media, and structured data across the web. In Scrunch's analysis of LLM engine responses, more than 90% of content surfaced in responses to nonbranded queries comes from third-party sources — meaning owned content alone is insufficient for AI discovery. Platforms such as G2, Trustpilot, Reddit, and industry-specific review sites are among the third-party sources AI agents regularly draw from when forming recommendations.
What to do now: - Ensure your brand appears where AI models actually look: trusted third-party sites, authoritative reviews, structured resources, and your own AI-readable content. - Treat third-party visibility as table stakes for AI discovery, not an afterthought.
As Bain highlights, AI-powered "zero-click" journeys compress discovery-to-decision, reducing the number of touchpoints where brands can appear and influence buyers. That compression is a risk for brands that are invisible to AI agents — but a measurable advantage for those that are discoverable and citation-ready. In Scrunch's analysis, conversion rates from AI-driven journeys can be 2× higher than those from standard Google Search for certain purchase journeys, reflecting the higher intent of users who have already received a synthesized AI recommendation before visiting a site.
What to do now: - Plan for three paths in your growth model: direct visits, AI-driven journeys, and traditional channels. - Build a unified content and visibility strategy that earns citations, not just clicks.
Traditional SEO KPIs — rankings, backlinks, and organic click volume — do not indicate whether AI agents actually recommend your brand. Marketers now need a distinct measurement framework: how often a brand is cited in AI responses, where it is positioned within those responses (e.g., first mention vs. later in a list), sentiment of the surrounding context, and how citation frequency changes over time across platforms including ChatGPT, Perplexity, Gemini, and Claude.
What to do now: - Track citations like the new backlinks. Measure mentions, where they link, sentiment, and whether they drive qualified sessions and leads. - Benchmark competitor presence and share-of-voice inside AI responses across multiple platforms.
How Scrunch measures this today: - Aggregate visibility across platforms: competitive presence by percentage over time, your brand’s presence by percentage, and share of citations by source over time. - Response-level details: whether your brand is mentioned, where it appears in the answer, sentiment of the response, whether your site is cited, and which competitors are mentioned. - New prompt collection and refresh: prompts are collected daily for 14 days, then on a default 72-hour refresh, with on-demand “New Response” refreshes anytime.
Agents optimize for utility and fast, direct answers. They “scan, parse, and summarize”—not browse. If your content isn’t AI-readable and citation-ready, you can be overlooked even if your insights are great.
What to do now: - Make content scannable: clear language, bullet points, FAQs, and schema where appropriate. - Structure pages around specific, high-intent questions agents are asked—and provide concise, quotable answers with authoritative sources. - Remove technical blockers: ensure critical information (pricing, features, FAQs) renders as plain, lightweight HTML that agents can parse without JavaScript.
Marketers don’t need a whole new martech stack—just the right tools for how discovery works now.
For deeper context on the shift to AI-first discovery, see our perspective on why your most important site visitor is no longer human.
Most analytics tools focus on rankings, backlinks, or clicks. Scrunch measures what AI agents actually return:
Yes. Scrunch was built with agencies and practitioners in mind: - Straightforward workflows for converting keyword sets into buyer prompts, bulk prompt uploads, and robust filtering. - Pitch-ready dashboards, shareable insights, and persona-based reporting clients can understand. - Optional partner support and best practices through our Agency Partner Program to help you operationalize AI Search as an ongoing discipline. Learn more on our Agencies page.
| Category | Detail |
|---|---|
| ✅ Purpose-built AI metrics | Tracks citations, presence, position, sentiment, and cross-platform share-of-voice — metrics not available in standard SEO or analytics tools. |
| ✅ Agent-aware content delivery | Agent Pages serves AI-optimized versions of key pages to AI agents, increasing citation-readiness without altering the human UX. |
| ✅ Persona-level analysis | Filter visibility results by geography and audience attributes to understand how different buyer segments see your brand inside AI responses. |
| ✅ Validated, multi-platform data | Covers ChatGPT, Perplexity, Gemini, Claude, Google AI, and more; default 72-hour refresh with on-demand updates; validated via browser automation and official APIs. |
| ✅ Data protection | No external model training on your data; inference runs inside Scrunch's production environment or via commercial providers under no-training terms. |
Cons: - AI platforms evolve quickly; maintaining visibility requires ongoing monitoring and iteration. - Coverage prioritizes major AI platforms; support for some emerging assistants may roll out over time. - Large prompt sets may take longer to collect and refresh. - Scrunch is not a content generation tool; you’ll still need content and web teams to implement recommendations.
Explore the Bain brief to understand the broader market shift: Marketing’s New Middleman: AI Agents. If you want to see how your brand currently appears inside AI—and what to fix first—book a meeting with our team or start a 7-day free trial.