How is AI agent or Muse affecting companies like EverQuote, which are like insurance marketplace? Give your answers with data, like hard data points. And then what has EverQuote's management and other similar marketplaces companies management said about agent and how it will affect their business, etc., and how much their traffic currently or revenue is coming from agent? Also, tell me about other marketplaces that might be affected by Muse or agentic AI.

Autonomous AI agents—most notably Meta’s personal AI agent, Muse, alongside OpenAI’s Operator/ChatGPT ecosystem and Google’s AI Overviews—have created a sharp divide between public equity sentiment and fundamental operational performance for insurance marketplaces like EverQuote (NASDAQ: EVER), QuinStreet (NASDAQ: QNST), MediaAlpha (NYSE: MAX), and LendingTree (NASDAQ: TREE).

While Wall Street has priced in fears of top-of-funnel disintermediation (compressing EverQuote’s valuation to ~6.7x P/E), actual fundamental data and third-party testing show that consumer-facing AI agents struggle to bypass insurance marketplaces due to strict state-level regulatory compliance, complex actuarial underwriting, and carrier rate opacity. Instead, marketplaces are using internal AI bidding engines and LLM ad integrations to post record revenues and margin expansion.

1. Hard Data Points: How Muse & AI Agents Are Affecting EverQuote and Peers

The impact of agentic AI on insurance marketplaces breaks down into three distinct data buckets: equity market reaction, real-world agent testing, and financial/operational efficiency.

The "Muse" Selloff & Valuation Compression (Sept–Oct 2026)

• The Trigger: After Meta’s autonomous personal AI agent, Muse, hit #1 on the US App Store on September 18, 2026—demonstrating the ability to autonomously browse the web on a secure cloud computer, book flights, shop, and fill out forms—marketplace stocks suffered a sharp multi-day selloff driven by fears of "disintermediation of direct user gateways".
• Stock Impact: On September 22–23, 2026, EverQuote (EVER) fell 8.2% in a single session (sliding ~16% total from its August levels to a low of $19.02 before rebounding to ~$22–$23 in late October). Peer marketplaces sold off in sympathy during the same window: Cars.com (CARS) dropped 5.7%, CarGurus (CARG) fell 5.1%, Etsy (ETSY) fell 4.1%, and Booking Holdings (BKNG) slipped 3.3%.

Autonomous Research Testing of Meta’s Muse in Insurance (Oct 20, 2026)

Financial research firm Autonomous Research published a benchmark study evaluating Meta’s Muse AI agents across insurance shopping and claims tasks:

• Quote Retrieval Failure: Autonomous found that Muse had "mixed results fetching real insurance quotes" and failed to outperform human agents or existing comparison funnels. Because P&C carriers guard rate tables behind multi-step walls and dynamic underwriting filters, autonomous web-scraping agents frequently get blocked or fail to bind policies.
• Where Muse Succeeded (and Created Carrier Risk): Once manually fed the right policy data, Muse performed well at summarizing coverage, pricing, and discounts. More importantly, Autonomous flagged that Muse actively reminds consumers of mandatory post-accident actions and connects them directly with personal injury law firms for auto claims—a dynamic researchers warned will drive AI-fueled tort inflation and higher loss ratios for carriers rather than killing lead-gen marketplaces.

Financial & Operational Performance Data

Despite disintermediation fears, P&C insurance marketplaces are posting record financial metrics, fueled by carriers shifting from rate-hardening back to growth mode (with major carrier combined ratios sitting in the healthy 80s):

2. Traffic & Revenue Currently Coming from AI Agents

If you look strictly at inbound consumer traffic and revenue originating from third-party LLM agents (ChatGPT, Claude, Perplexity, Meta Muse), the current contribution for EverQuote and its peers is de minimis (well under 1% of total volume), though growing rapidly from a near-zero base.

• Third-Party Traffic Analytics (August/September 2026): According to Semrush LLM traffic tracking for everquote.com, identifiable direct referral traffic from AI platforms remains tiny relative to its millions of paid auction clicks—registering roughly 7,000 visits from specialized AI matchers (insurematch.ai) and ~250 direct monthly visits from ChatGPT.
• Why LLM Traffic Is Incremental to EverQuote: On EverQuote’s Q1 2026 earnings call, CEO Jayme Mendal explicitly pointed out a crucial structural nuance: "Right now we have no SEO traffic, right? We're historically paid." Unlike content-heavy publishers (e.g., NerdWallet or Bankrate) whose organic SEO traffic is cannibalized when LLMs answer questions directly, EverQuote acquires virtually 100% of its users via paid performance auctions on Google, Meta, and TikTok. Therefore, any traffic captured from LLM apps or AI chat ads is 100% incremental rather than cannibalistic.
• First Revenues Recorded by Peers: While neither EverQuote nor QuinStreet breaks out exact dollar figures for LLM-originated revenue in their SEC segment footnotes yet, QuinStreet confirmed in May and August 2026 that it is already live on OpenAI’s advertising platform across both insurance and home services and has "already generated first revenues from OpenAI’s platform."

3. What Marketplace Management Teams Are Saying

Across 2026 earnings calls, insurance marketplace executives have been heavily questioned by analysts (such as Maria Ripps at Canaccord and Jason Kreyer at Craig-Hallum) regarding LLM traffic, ChatGPT apps, and agentic disintermediation.

EverQuote (CEO Jayme Mendal & CFO Joseph Sanborn)

• On ChatGPT Apps & Current Volume (Q1 & Q2 2026 Calls): Mendal noted that EverQuote outlines three pillars to capture LLM traffic: content integrations, native AI apps, and paid LLM ads. In Q1, Mendal expressed skepticism about early clones: "There's probably 15-plus apps in the insurance category right now in ChatGPT. They're all quite similar. They kind of just take a web form experience and apply a lightweight conversational front end to it before spinning the consumer back to a web-based quoting and binding experience. Our sense is there's very little actual traffic flowing through those at this point." By the Q2 call (August 2026), Mendal confirmed: "We launched our ChatGPT app last quarter, and it's getting some traffic," while emphasizing that the real scale will come as ChatGPT and other LLMs open up paid advertising inventory, where EverQuote can leverage its position as one of the largest paid advertisers in insurance.
• On "Agentic Traffic Operations" & Token Economics (Q2 2026 Call): Internally, Mendal explained that EverQuote is transitioning its bidding platforms toward "agentic traffic operations"—using AI agents to simulate consumer personas and autonomously manage massive traffic bidding portfolios with minimal human intervention. CFO Joseph Sanborn added that while internal LLM token usage is ramping up, "we have religion around efficiency. And so we have good visibility on the token usage and the expenses associated with it... we've doubled revenue over the last couple of years while keeping OpEx flat."
• The WaniWani Partnership (August 10, 2026): Immediately following Q2 earnings, EverQuote announced an exclusive commercial partnership and minority equity investment in WaniWani, an agentic AI distribution and compliance infrastructure startup (founded by former Luko CEO Raphaël Vullierme). WaniWani builds Model Context Protocol (MCP) services and synthetic buyer simulations that allow P&C insurers and EverQuote’s 5,000+ local agents to surface real-time, compliant quotes directly inside ChatGPT, Claude, Gemini, and Meta’s Muse, complete with audit trails for GDPR, SOC 2, and state insurance disclosures.

QuinStreet (CEO Doug Valenti)

• On Google AI Overviews & OpenAI Ads (Q3 & Q4 FY26 Calls): Valenti directly refuted the bear thesis that AI search kills marketplace lead generation: "As AI Overviews have expanded rapidly over the past year to now trigger on an estimated 50% plus of Google searches, revenue from our proprietary campaigns on Google has grown by over 100% over the same period. We are an early participant in OpenAI's advertising platform, where we are live with campaigns in both insurance and home services."
• Why AI Agents Can't Easily Bypass Marketplaces: Valenti argued that insurance marketplaces sit on the complex plumbing—integrating dynamic carrier rate tables, pre-qualifying contact center leads, and matching carrier underwriting appetite in real time—making them essential monetization partners for LLM platforms rather than targets for replacement.

MediaAlpha (CEO Steve Yi)

• On Higher-Intent Conversions from LLM Search (Q2 2026 Call, July 29, 2026): Steve Yi highlighted that consumers arriving via conversational AI searches actually convert better and carry higher value per lead: "What they're doing with an LLM search is that they're going deeper and expressing more nuances and more details around the insurance that they're looking for. And so what you have is a more targeted consumer." MediaAlpha posted record Q2 revenue of $316.9 million (+26% YoY), noting that carriers #3, #4, and #5 on its platform grew spend by 300% YoY as they leaned into digital customer acquisition.

4. Other Marketplaces Vulnerable to (or Reshaped by) Muse & Agentic AI

The September 2026 "Muse selloff" revealed exactly how institutional investors categorize marketplace risk. Whether an autonomous agent like Muse disrupts or empowers a marketplace depends on regulatory friction, inventory fragmentation, and revenue model (SEO vs. Paid/API):

Want to dig deeper into EverQuote's unit economics heading into their Q3 earnings report on November 2, or examine how WaniWani's MCP infrastructure actually integrates carrier bidding into LLMs?