JayOh University / Daily Intel / Saturday, October 31, 2026

Google's AI Mode Is Eating Its Own Revenue

  • 6 insights
  • 4 sources
  • ~4 min read
  • GTM · RevOps · AI

The 60-second scan

Click any line for the full breakdown
GTM & RevOps
01Latent SpaceGoogle's CBO linked AI Mode's lower monetization to a ~15%, $500B single-day drop, the worst since its 2004 IPO.start here→ 02PPC LandGoogle and Bing will honor one shared robots.txt rule to block AI training, as a study puts AI Overviews CTR at just 0.9%.→
Platform & Tool Releases
03HubSpotSuper Admins can now let ChatGPT, Claude and Copilot read Sensitive Data properties, read-only, in public beta.→ 04HubSpotCommerce Hub Pro/Enterprise customers get a connected quote-to-cash flow, one quote can now generate multiple invoices.→ 05Mistral AIMistral open-sourced a trillion-parameter model built for self-hosted deployment and cybersecurity use cases.→ 06HubSpotBreeze Assistant can search a customer's full email and chat history and answer questions about recurring issues or objections.→

GTM & RevOps

01Latent Space

Google admits AI Mode is cannibalizing Search revenue

Google's CBO linked AI Mode's lower monetization to a ~15%, $500B single-day drop, the worst since its 2004 IPO.

What happened
  • Google Chief Business Officer Philipp Schindler cited "the lower monetization we were seeing in AI Mode" as a factor in the earnings miss.
  • Alphabet shares fell roughly 15% in a single day, a swing described as the worst since Google's 2004 IPO.
  • The drop set off a broader selloff across tech stocks tied to AI infrastructure spend.
Why it matters

If Google's own numbers show AI answers monetize worse than search ads, the paid-search channel B2B teams have leaned on for demand gen is entering a structurally lower-yield phase.

The JayOh take →

Pressure-test paid search ROI assumptions in 2027 planning now, and shift incremental budget toward channels less exposed to AI-answer cannibalization: direct outbound, partner/affiliate, and owned communities.

Source: Latent SpaceBack to the scan ↑

02PPC Land

Google and Bing back a shared AI-training opt-out signal

Google and Bing will honor one shared robots.txt rule to block AI training, as a study puts AI Overviews CTR at just 0.9%.

What happened
  • Google and Bing agreed to honor a single IETF-track robots.txt signal for opting content out of AI training.
  • It replaces fragmented, crawler-specific rules like Google-Extended, GPTBot, and ClaudeBot with one shared standard.
  • A cited study measured AI Overviews click-through at just 0.9%.
  • Bing is separately testing ads inside its AI answer surfaces.
Why it matters

Revenue teams finally get one lever to control how AI models use their content, and the sub-1% AI Overviews CTR is hard evidence that owned distribution still outperforms chasing search visibility.

The JayOh take →

Set the new unified AI-training opt-out at the domain level as a default governance step, but don't treat it as a traffic fix, keep investing in direct channels since AI Overviews aren't sending meaningful click volume regardless.

Source: PPC LandBack to the scan ↑

Platform & Tool Releases

03HubSpot

HubSpot opens Sensitive Data to AI connectors (beta)

Super Admins can now let ChatGPT, Claude and Copilot read Sensitive Data properties, read-only, in public beta.

What happened
  • ChatGPT, Claude and Copilot can access Sensitive Data properties and engagements once a Super Admin opts in.
  • Gemini and the Remote MCP server get engagement access only, not Sensitive Data properties.
  • Access is strictly read-only, and Highly Sensitive Data stays restricted.
  • New permissions can take up to two hours to appear after reconnecting a connector.
Why it matters

This closes a real gap for teams using AI connectors for deal or account research who were previously blocked from sensitive fields, but it also widens the blast radius of a connector misconfiguration.

The JayOh take →

Before opting in, audit which properties are marked Sensitive Data and confirm only the connectors and users who truly need that access get it, this is a permissions review to run proactively, not after an incident.

Source: HubSpotBack to the scan ↑

04HubSpot

HubSpot's connected CPQ, billing and payments reaches GA

Commerce Hub Pro/Enterprise customers get a connected quote-to-cash flow, one quote can now generate multiple invoices.

What happened
  • Connected CPQ, Billing, and Payments is now generally available for Commerce Hub Professional and Enterprise.
  • Restrictions on billing schedules are removed, and a single quote can generate multiple payable invoices.
  • Buyers can be guided to set up payment directly when they accept a quote.
Why it matters

This closes a long-standing quote-to-cash gap in HubSpot, letting revenue teams run multi-invoice billing, retainers, milestones, usage, natively instead of bolting on a separate billing tool.

The JayOh take →

For clients already live on Commerce Hub Pro/Enterprise, this is worth a scoping call now, it can retire a third-party billing integration and simplify the deal-to-cash handoff that otherwise gets custom-built.

Source: HubSpotBack to the scan ↑

05Mistral AI

Mistral releases open weights for 1T-parameter Large 4

Mistral open-sourced a trillion-parameter model built for self-hosted deployment and cybersecurity use cases.

What happened
  • Mistral released open weights for Mistral Large 4 ("Le Chonk"), a roughly 1 trillion parameter multimodal model.
  • Positioned as the best open-weight model for coding, trained in Europe.
  • Deployable on self-managed infrastructure via vLLM or SGLang, under a modified MIT (Mistral Ecosystem Open-Weight) license.
Why it matters

A credible, self-hostable frontier-scale model gives data-sensitive B2B companies, regulated industries and PE-backed targets with strict data residency rules, a real alternative to API-only frontier models.

The JayOh take →

Keep this on the radar for clients who've ruled out AI features over data residency or vendor lock-in concerns, a self-hosted option changes that conversation, even though it requires real infra investment to run well.

Source: Mistral AIBack to the scan ↑

Catch up on past briefings in JayOh University.

Turn these signals into your GTM roadmap.

We engineer the CRM, RevOps and AI systems behind B2B revenue teams.

Talk to JayOh