Marcus Thibodeaux

A framework for measuring how well DSPs score ad intent from conversational context.

LLM advertising needs auctions over output distributions, not impression slots.

Publishers must evaluate AI ad partners on conversational fit, not web metrics.

Conversational ads need live context matching, not page-based targeting.

Advertisers must reach buyers inside AI conversations before they make decisions.

ChatGPT ads are buyable through three paths; Perplexity's are not.

Ads in AI chat fire on conversation context, not keywords.

Advertisers must match ads to live conversation context, not keywords.

Platforms trade conversational depth for cross-surface reach in different ways.

Separating true conversational AI ads from three unrelated concepts costing marketers real money.

Agencies must now build audit trails into AI ad campaigns, not assemble them after regulators ask.

Centralized accounts pool volume but create audit risk as AI platform policies shift monthly.

Programmatic buyers need new mental models to bid on ads inside AI conversations.

Different prompt types signal different purchase stages, not just different keywords.

AI systems can now read unspoken buyer intent hidden within conversation patterns.

Conversational AI gives advertisers direct access to purchase intent, not guesses.

Prompt data captures purchase intent before the first click, revealing why customers buy.